Ocular surface interferometry (osi) apparatus for imaging an ocular tear film

<u>Field of the Disclosure</u>

0001The technology of the disclosure relates to imaging an ocular tear film. The technology of the disclosure also relates to measuring ocular tear film layer thickness(es), including lipid layer thickness (LLT) and/or aqueous layer thickness (ALT). Imaging the ocular tear film and measuring TFLT may be used to diagnose "dry eye," which may be due to any number of deficiencies, including lipid deficiency and aqueous deficiency.

<u>Background</u>

0002In the human eye, the precorneal tear film covering ocular surfaces is composed of three primary layers: the mucin layer, the aqueous layer, and the lipid layer. Each layer plays a role in the protection and lubrication of the eye and thus affects dryness of the eye or lack thereof. Dryness of the eye is a recognized ocular disease, which is generally referred to as "dry eye," "dry eye syndrome" (DES), or "keratoconjunctivitis sicca" (KCS). Dry eye can cause symptoms, such as itchiness, burning, and irritation, which can result in discomfort. There is a correlation between the ocular tear film layer thicknesses and dry eye disease. The various different medical conditions and damage to the eye as well as the relationship of the aqueous and lipid layers to those conditions are reviewed in <nplcit id="ncit0001" npl-type="s"><text>Surv Opthalmol 52:369-374, 2007</text></nplcit> and additionally briefly discussed below.

0003As illustrated in <figref idref="f0001">Figure 1</figref>, the precorneal tear film includes an innermost layer of the tear film in contact with a cornea 10 of an eye 11 known as the mucus layer 12. The mucus layer 12 is comprised of many mucins. The mucins serve to retain aqueous in the middle layer of the tear film known as the aqueous layer. Thus, the mucus layer 12 is important in that it assists in the retention of aqueous on the cornea 10 to provide a protective layer and lubrication, which prevents dryness of the eye 11.

0004A middle or aqueous layer 14 comprises the bulk of the tear film. The aqueous layer 14 is formed by secretion of aqueous by lacrimal glands 16 and accessory tear glands 17 surrounding the eye 11, as illustrated in <figref idref="f0001">Figure 2</figref>. The aqueous, secreted by the lacrimal glands 16 and accessory tear glands 17, is also commonly referred to as "tears." One function of the aqueous layer 14 is to help flush out any dust, debris, or foreign objects that may get into the eye 11. Another important function of the aqueous layer 14 is to provide a protective layer and lubrication to the eye 11 to keep it moist and comfortable. Defects that cause a lack of sufficient aqueous in the aqueous layer 14, also known as "aqueous deficiency," are a common cause of dry eye. Contact lens wear can also contribute to dry eye. A contact lens can disrupt the natural tear film and can reduce corneal sensitivity over time, which can cause a reduction in tear production.

0005The outermost layer of the tear film, known as the "lipid layer" 18 and also illustrated in <figref idref="f0001">Figure 1</figref>, also aids to prevent dryness of the eye. The lipid layer 18 is comprised of many lipids known as "meibum" or "sebum" that is produced by meibomian glands 20 in upper and lower eyelids 22, 24, as illustrated in <figref idref="f0002">Figure 3</figref>. This outermost lipid layer is very thin, typically less than 250 nanometers (nm) in thickness. The lipid layer 18 provides a protective coating over the aqueous layer 14 to limit the rate at which the aqueous layer 14 evaporates. Blinking causes the upper eyelid 22 to mall up aqueous and lipids as a tear film, thus forming a protective coating over the eye 11. A higher rate of evaporation of the aqueous layer 14 can cause dryness of the eye. Thus, if the lipid layer 18 is not sufficient to limit the rate of evaporation of the aqueous layer 14, dryness of the eye may result.

0006Notwithstanding the foregoing, it has been a long standing and vexing problem for clinicians and scientists to quantify the lipid and aqueous layers and any deficiencies of same to diagnose evaporative tear loss and/or tear deficiency dry eye conditions. Further, many promising treatments for dry eye have failed to receive approval from the United States Food and Drug Administration due to the inability to demonstrate clinical effectiveness to the satisfaction of the agency. Many clinicians diagnose dry eye based on patient symptoms alone. Questionnaires have been used in this regard. Although it seems reasonable to diagnose dry eye based on symptoms alone, symptoms of ocular discomfort represent only one aspect of "dry eyes," as defined by the National Eye Institute workshop on dry eyes. In the absence of a demonstrable diagnosis of tear deficiency or a possibility of excessive tear evaporation and damage to the exposed surface of the eye, one cannot really satisfy the requirements of dry eye diagnosis.

0007<patcit id="pcit0001" dnum="US20080319323A"><text>US 2008/0319323</text></patcit> describes methods and apparatus for tear film measurement, and in particular measurement of the thickness of the lipid layer component of the precorneal tear film on the surface subsequent to blinking. Light is directed to the lipid layer of a patient's eye by use of a broad spectrum, large area Lambertian light source which produces specularly reflected light rays. These rays undergo constructive and destructive interference in the lipid layer. The specularly reflected light is collected and focused such that the interference patterns on the tear film lipid layer are observable. For detailed visualization and good digital imaging, a CCD video camera having a resolution of up to 1280 × 1024 pixels and at least 15 Hz frame rate is suggested, to show the progression of lipid interference patterns as they spread across the eye. An output signal is produced that is representative of the specularly reflected light which is suitable for further analysis, such as projection on to a high resolution video monitor or analysis by or storage in a computer. The desirability of enhancing contrast and thus the observability and measurability of the lipid layer thickness of the precorneal tear film is mentioned. Given the data bandwidth, a high performance computer system is said to be needed. Software to operate the camera, capture the images, store and retrieve image files, and execute chosen calculations on the data is said to be critical, and in some cases the system is fully motorized and operated in manual, semi-autonomous, or autonomous modes, depending upon the control software. A folly automatic system is said to adjust the mechanical aspects of the system, focus the camera, record the motion of the lipid film, calculate various measurements of the film structure, report an assessment of the quality of the lipid film, and record the data in the patient's record file.

0008<patcit id="pcit0002" dnum="US6088470A"><text>US patent no. 6,088,470</text></patcit> describes methods for method and apparatus for removal of bright or dark spots by the fusion of multiple images. The document acknowledges that known method of achieving these objectives can be time-consuming. It sets out to provide an imaging system that can obtain useful images of the eye while minimizing the effect of bright spots, i.e. specular reflections caused by the system's own illumination, without requiring the subject to remove any eyeglasses or contact lenses that may be present. It is suggested to combine camera and lighting control with a particular image fusion technique to produce high quality images that are free of bright spots or specularities. Multiple light sources are selected with relatively wide spacing from one another. Images of the subject captured in quick succession and illuminated by different light sources are fused. The images are taken or modified to pair pixels one from each image such that both pixels in the pair correspond to a single point on the subject in the image. To create an image free of specular reflections the pixel from each pair having the lower gray scale value is selected. Shadows can be removed by selecting the pixel having the higher gray scale value only if that value is below a threshold.

<u>Summary of the Detailed Description</u>

0009Embodiments of the detailed description include ocular surface interferometry (OSI) devices, systems, and methods for imaging an ocular tear film and/or measuring a tear film layer thickness (TFLT) in a patient's ocular tear film. The OSI devices, systems, and methods can be used to measure the thickness of the lipid layer component (LLT) and/or the aqueous layer component (ALT) of the ocular tear film. "TFLT" as used herein includes LLT, ALT, or both LLT and ALT. "Measuring TFLT" as used herein includes measuring LLT, ALT, or both LLT and ALT. Imaging the ocular tear film and measuring TFLT can be used in the diagnosis of a patient's tear film, including but not limited to lipid layer and aqueous layer deficiencies. These characteristics may be the cause or contributing factor to a patient experiencing dry eye syndrome (DES).

0010In this regard, embodiments disclosed herein include a light source that is controlled to direct light in the visible region to an ocular tear film. The light source may be a Lambertian emitter that provides a uniform or substantially uniform intensity in all directions of emission. The light source is arranged such that light rays emitted from the light source are specularly reflected from the tear film and undergo constructive and destructive optical wave interference interactions (also referred to as "interference interactions") in the ocular tear film. An imaging device having a detection spectrum that includes the spectrum of the light source is focused on an area(s) of interest on the lipid layer of the tear film. The imaging device captures the interference interactions (i.e., modulation) of specularly reflected light rays from the illuminated tear film coming together by the focusing action of the imaging device in a first image. The imaging device then captures the optical wave interference signals (also referred to as "interference signals") representing the interference interactions of specularly reflected light from the tear film. The imaging device produces an output signal(s) representative of the interference signal in a first image. The first image may contain an interference signal for a given imaged pixel or pixels of the lipid layer by the imaging device.

0011The first image can be displayed to a technician or other user. The first image can also be processed and analyzed to measure a TFLT in the area or region of interest of the ocular tear film. In one embodiment, the first image also contains a background signal(s) that does not represent specularly reflected light from the tear film which is superimposed on the interference signal(s). The first image is processed to subtract or substantially subtract out the background signal(s) superimposed upon the interference signal to reduce error before being analyzed to measure TFLT. This is referred to as "background subtraction" in the present disclosure. The separate background signal(s) includes returned captured light that is not specularly reflected from the tear film and thus does not contain optical wave interference information (also referred to as "interference information"). For example, the background signal(s) may include stray, ambient light entering into the imaging device, scattered light from the patient's face and eye structures outside and within the tear film as a result of ambient light and diffuse illumination by the light source, and eye structure beneath the tear film, and particularly contribution from the extended area of the source itself. The background signal(s) adds a bias (i.e., offset) error to the interference signal(s) thereby reducing interference signal strength and contrast. This error can adversely influence measurement of TFLT. Further, if the background signal(s) has a color hue different from the light of the light source, a color shift can also occur to the captured optical wave interference (also referred to as "interference") of specularly reflected light thus introducing further error.

0012In this regard, the invention is set out in appended claim 1 and details of its implementation as reflected in the embodiments described herein are set out in the dependent claims.

0013In this regard, an optically "tiled" or "tiling" illumination of the tear film is provided. Tiling involves spatially controlling a light source to form specific lighting patterns on the light source when illuminating a portion(s) in an area or region of interest on the tear film in a first mode to obtain specularly reflected light and background signal(s). In this embodiment, the background signal(s) in the second image additionally includes scattered light as a result of diffuse illumination by the light source. Because background signal(s) due to scattered light as a result of diffuse illumination by the light source is also present in the first image, capturing a second image that includes diffuse illumination by the light source can further reduce bias (i.e., offset) error and increase interference signal strength and contrast over embodiments that do not control the light source to illuminate the tear film when the second image is captured.

0014In this regard, and in accordance with the invention as set out in the claims, the light source is controlled in a first mode to provide a lighting pattern to produce specularly reflected light from a first portion(s) in the area or region of interest of the tear film while obliquely illuminating an adjacent, second portion(s) of the area or region of interest of the tear film. The imaging device captures a first image representing the interference of the specularly reflected light with additive background signal(s) from the first portion(s) of the area or region of interest, and background signal(s) from a second portion(s) of the area or region of interest. The background signal(s) from the second portion(s) includes scattered light as a result of diffuse reflection of the illumination by the light source, and ambient light. The light source is then alternately controlled in a second mode to reverse the lighting pattern of the first mode to capture specularly reflected light from the second portion(s) in the area or region of interest of the tear film while obliquely illuminating the first portion(s) in the area or region of interest of the tear film. The imaging device captures a second image representing the interference of the specularly reflected light and with additive background signal(s) from the second portion(s) in the area or region of interest on the tear film, and background signal(s) from the first portion(s) in the area or region of interest on the tear film. The background signal(s) from the first portion(s) includes scattered light as a result of diffuse reflection of the illumination by the light source. The first and second images are combined to subtract or substantially subtract background offset from the interference signals to produce the resulting image. Again, the resulting image can be displayed on a visually display to be analyzed by a technician and processed and analyzed to measure a TFLT.

0015After the interference of the specularly reflected light is captured and a resulting image containing the interference signal is produced from any method or device disclosed in this disclosure, the resulting image can also be pre-processed before being processed and analyzed to measure TFLT. Pre-processing can involve performing a variety of methods to improve the quality of the resulting signal, including but not limited to detecting and removing eye blinks or other signals in the captured images that hinder or are not related to the tear film. After pre-processing, the interference signal or representations thereof can be processed to be compared against a tear film layer interference model to measure TFLT. The interference signal can be processed and converted by the imaging device into digital red-green-blue (RGB) component values which can be compared to RGB component values in a tear film interference model to measure TFLT on an image pixel-by-pixel basis. The tear film interference model is based on modeling the lipid layer of the tear film in various thicknesses and mathematically or empirically observing and recording resulting interference interactions of specularly reflected light from the tear film model when illuminated by the light source and detected by a camera (imaging device).

0016In a tear film interference model, the lipid layer is modeled of various LLTs to observe interference interactions resulting from the various LLTs. The aqueous layer may be modeled in the tear film interference model to be of an infinite, minimum, or varying thickness. If the aqueous layer is modeled to be of an infinite thickness, the tear film interference model assumes no specular reflections occur from the aqueous-to-mucin layer transition. If the aqueous layer is modeled to be of a certain minimum thickness (∼>2µm e.g.), the effect of specular reflection from the aqueous-to-mucin layer transition may be considered in the resulting interference. In either case, the tear film interference model is a 2-wave tear film interference model to represent the interference between specularly reflected light from the air-to lipid layer transition and the lipid-to-aqueous layer transition. Thus, a 2-wave tear film interference model will include one-dimension of data comprised of interference interactions corresponding to the various LLTs. In this case, to measure LLT the interference interactions in the interference signal representing specularly reflected light from the tear film produced by the imaging device are compared to the interference patterns in the tear film interference model. However, if the aqueous layer is also modeled to be of varying ALTs, the tear film interference model will be a 3-wave tear film interference model. The 3-wave tear film interference model will include interference between the air-to lipid layer, lipid-to-aqueous layer, and aqueous-to-mucus/cornea layer transitions. As a result, a 3-wave tear film interference model will include two-dimensions of data comprised of interference interactions corresponding to various LLT and ALT combinations. In this case, to measure LLT and/or ALT the interference interactions from the interference signal representing specularly reflected light from the tear film produced by the imaging device can be compared to interference interactions in the 3-wave tear film interference model.

0017The tear film interference model can be a theoretical tear film interference model where the light source and the tear film layers are modeled mathematically. The tear film layers may be mathematically modeled by modeling the tear film layers after certain biological materials. Interference interactions from the mathematically modeled light source illuminating the mathematically modeled tear film and received by the mathematically modeled camera are calculated and recorded for varying TFLTs. Alternatively, the tear film interference model can be based on a biological or phantom tear film model comprised of biological or phantom tear film layers. The actual light source is used to illuminate the biological or phantom tear film model and interference interactions representing interference of specularly reflected light are empirically observed and recorded for various TFLTs using the actual camera.

0018Those skilled in the art will appreciate the scope of the present invention and realize additional aspects thereof after reading the following detailed description of the preferred embodiments in association with the accompanying drawing figures.

<u>Brief Description of the Drawings</u>

0019The accompanying drawing figures incorporated in and forming a part of this specification illustrate several aspects of the invention, and together with the description serve to explain the principles of the invention.

0020The invention as claimed is depicted in and described with reference to <figref idref="f0054">figures 11 - 53</figref>. The remaining figures are included for background and context. <ul id="ul0001" list-style="none" compact="compact"><li><figref idref="f0001">Figure 1</figref> is a side view of an exemplary eye showing the three layers of the tear film in exaggerated form;</li><li><figref idref="f0001">Figure 2</figref> is a front view of an exemplary eye showing the lacrimal and accessory tear glands that produce aqueous in the eye;</li><li><figref idref="f0002">Figure 3</figref> illustrates exemplary upper and lower eyelids showing the meibomian glands contained therein;</li><li><figref idref="f0003">Figures 4A</figref> and <figref idref="f0004">4B</figref> are illustrations of an exemplary light source and imaging device to facilitate discussion of illumination of the tear film and capture of interference interactions of specularly reflected light from the tear film;</li><li><figref idref="f0005">Figure 5</figref> illustrates (in a microscopic section view) exemplary tear film layers to illustrate how light rays can specularly reflect from various tear film layer transitions;</li><li><figref idref="f0006">Figure 6</figref> is a flowchart of an exemplary process for obtaining one or more interference signals from images of a tear film representing specularly reflected light from the tear film with background signal subtracted or substantially subtracted;</li><li><figref idref="f0007">Figure 7</figref> illustrates a first image focused on a lipid layer of a tear film and capturing interference interactions of specularly reflected light from an area or region of interest of the tear film;</li><li><figref idref="f0008">Figure 8</figref> illustrates a second image focused on the lipid layer of the tear film in <figref idref="f0007">Figure 7</figref> and capturing background signal when not illuminated by the light source;</li><li><figref idref="f0009">Figure 9</figref> illustrates an image of the tear film when background signal captured in the second image of <figref idref="f0008">Figure 8</figref> is subtracted from the first image of <figref idref="f0007">Figure 7</figref>;</li><li><figref idref="f0010">Figure 10</figref> is a flowchart of another exemplary optical tiling process for obtaining one or more interference signals from tiled portions in an area or region of interest of a tear film representing specularly reflected light from the tear film with background signal subtracted or substantially subtracted;</li><li><figref idref="f0011">Figure 11A</figref> illustrates a first image focused on the lipid layer of the tear film capturing interference interactions of specularly reflected light and background signal from tiled portions in an area or region of interest of the tear film;</li><li><figref idref="f0011">Figure 11B</figref> illustrates a second image focused on the lipid layer of the tear film in <figref idref="f0011">Figure 11A</figref> capturing background signal and interference interactions of specularly reflected light from the tiled portions in the area or region of interest in <figref idref="f0011">Figure 11A</figref>, respectively;</li><li><figref idref="f0012">Figure 12</figref> illustrates an image when the background signal captured in diffusely illuminated tiled portions in the first and second images of <figref idref="f0011">Figures 11A and 11B</figref> are subtracted or substantially subtracted from the specularly reflected light in corresponding tiled portions in the first and second images of <figref idref="f0011">Figures 11A and 11B</figref>;</li><li><figref idref="f0013">Figure 13A</figref> illustrates a first image focused on a lipid layer of a tear film capturing interference interactions of specularly reflected light and background signal from concentric tiled portions in an area or region of interest of the tear film;</li><li><figref idref="f0013">Figure 13B</figref> illustrates a second image focused on a lipid layer of the tear film in <figref idref="f0013">Figure 13A</figref> capturing interference interactions of background signal and specularly reflected light, respectively, from the concentric tiled portions in the area or region of interest of the tear film in <figref idref="f0013">Figure 13A</figref>;</li><li><figref idref="f0014">Figure 14</figref> is a perspective view of an exemplary ocular surface interferometry (OSI) device for illuminating and imaging a patient's tear film, displaying images, analyzing the patient's tear film, and generating results from the analysis of the patient's tear film;</li><li><figref idref="f0015">Figure 15</figref> is a side view of the OSI device of <figref idref="f0014">Figure 14</figref> illuminating and imaging a patient's eye and tear film;</li><li><figref idref="f0016">Figure 16</figref> is a side view of a video camera and illuminator within the OSI device of <figref idref="f0014">Figure 14</figref> imaging a patient's eye and tear film;</li><li><figref idref="f0017">Figure 17</figref> is a top view of an illumination device provided in the OSI device of <figref idref="f0014">Figure 14</figref> illuminating a patient's tear film with the video camera capturing images of the patient's tear film;</li><li><figref idref="f0018">Figure 18</figref> is a perspective view of an exemplary printed circuit board (PCB) with a plurality of light emitting diodes (LED) provided in the illumination device of the OSI device in <figref idref="f0014">Figure 14</figref> to illuminate the patient's tear film;</li><li><figref idref="f0019">Figure 19</figref> is a perspective view of the illumination device and housing in the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0020 f0021">Figures 20-24</figref> illustrate exemplary light grouping patterns for the illumination device of <figref idref="f0017">Figure 17</figref> that may be used to image tiled patterns of specularly reflected light from a tear film;</li><li><figref idref="f0022">Figure 25A</figref> illustrates an exemplary system diagram of a control system and supporting components in the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0023">Figure 25B</figref> is a flowchart illustrating an exemplary overall processing flow of the OSI device of <figref idref="f0014">Figure 14</figref> having systems components according to the exemplary system diagram of the OSI device in <figref idref="f0022">Figure 25A</figref>;</li><li><figref idref="f0024">Figure 26</figref> is a flowchart illustrating exemplary pre-processing steps performed on the combined first and second images of a patient's tear film before measuring tear film layer thickness (TFLT);</li><li><figref idref="f0025">Figure 27</figref> is an exemplary graphical user interface (GUI) for controlling imaging, pre-processing, and post-processing settings of the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0026">Figure 28</figref> illustrates an example of a subtracted image in an area or region of interest of a tear film containing specularly reflected light from the tear film overlaid on top of a background image of the tear film;</li><li><figref idref="f0027">Figures 29A and 29B</figref> illustrate exemplary threshold masks that may be used to provide a threshold function during pre-processing of a resulting image containing specularly reflected light from a patient's tear film;</li><li><figref idref="f0028">Figure 30</figref> illustrates an exemplary image of <figref idref="f0026">Figure 28</figref> after a threshold pre-processing function has been performed leaving interference of the specularly reflected light from the patient's tear film;</li><li><figref idref="f0029">Figure 31</figref> illustrates an exemplary image of the image of <figref idref="f0028">Figure 30</figref> after erode and dilate pre-processing functions have been performed on the image;</li><li><figref idref="f0030">Figure 32</figref> illustrates an exemplary histogram used to detect eye blinks and/or eye movements in captured images or frames of a tear film;</li><li><figref idref="f0031">Figure 33</figref> illustrates an exemplary process for loading an International Colour Consortium (ICC) profile and tear film interference model into the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0032">Figure 34</figref> illustrates a flowchart providing an exemplary visualization system process for displaying images of a patient's tear film on a display in the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0033 f0034 f0035">Figures 35A-35C</figref> illustrate exemplary images of a patient's tear film with a tiled pattern of interference interactions from specularly reflected light from the tear film displayed on a display;</li><li><figref idref="f0036">Figure 36</figref> illustrates an exemplary post-processing system that may be provided in the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0037">Figure 37A</figref> illustrates an exemplary 3-wave tear film interference model based on a 3-wave theoretical tear film model to correlate different observed interference color with different lipid layer thicknesses (LLTs) and aqueous layer thicknesses (ALTs);</li><li><figref idref="f0038">Figure 37B</figref> illustrates another exemplary 3-wave tear film interference model based on a 3-wave theoretical tear film model to correlate different observed interference color with different lipid layer thicknesses (LLTs) and aqueous layer thicknesses (ALTs);</li><li><figref idref="f0039">Figure 38</figref> is another representation of the 3-wave tear film interference model of <figref idref="f0037">Figure 37</figref> with normalization applied to each red-green-blue (RGB) color value individually;</li><li><figref idref="f0040">Figure 39</figref> is an exemplary histogram illustrating results of a comparison of interference interactions from the interference signal of specularly reflected light from a patient's tear film to the 3-wave tear film interference model of <figref idref="f0037">Figures 37</figref> and <figref idref="f0039">38</figref> for measuring TFLT of a patient's tear film;</li><li><figref idref="f0041">Figure 40</figref> is an exemplary histogram plot of distances in pixels between RGB color value representation of interference interactions from the interference signal of specularly reflected light from a patient's tear film and the nearest distance RGB color value in the 3-wave tear film interference model of <figref idref="f0037">Figures 37</figref> and <figref idref="f0039">38</figref>;</li><li><figref idref="f0042">Figure 41</figref> is an exemplary threshold mask used during pre-processing of the tear film images;</li><li><figref idref="f0043">Figure 42</figref> is an exemplary three-dimensional (3D) surface plot of the measured LLT and ALT thicknesses of a patient's tear film;</li><li><figref idref="f0044">Figure 43</figref> is an exemplary image representing interference interactions of specularly reflected light from a patient's tear film results window based on replacing a pixel in the tear film image with the closest matching RGB color value in the normalized 3-wave tear film interference model of <figref idref="f0039">Figure 38</figref>;</li><li><figref idref="f0045">Figure 44</figref> is an exemplary TFLT palette curve for a TFLT palette of LLTs plotted in RGB space for a given ALT in three-dimensional (3D) space;</li><li><figref idref="f0046">Figure 45</figref> is an exemplary TFLT palette curve for the TFLT palette of <figref idref="f0045">Figure 44</figref> with LLTs limited to a maximum LLT of 240 nm plotted in RGB space for a given ALT in three-dimensional (3D) space;</li><li><figref idref="f0047">Figure 46</figref> illustrates the TFLT palette curve of <figref idref="f0046">Figure 45</figref> with an acceptable distance to palette (ADP) filter shown to determine tear film pixel values having RGB values that correspond to ambiguous LLTs;</li><li><figref idref="f0048">Figure 47</figref> is an exemplary login screen to a user interface system for controlling and accessing the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0049">Figure 48</figref> illustrates an exemplary interface screen for accessing a patient database interface in the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0050">Figure 49</figref> illustrates a patient action control box for selecting to either capture new tear film images of a patient in the patient database or view past captured images of the patient from the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0051">Figure 50</figref> illustrates a viewing interface for viewing a patient's tear film either captured in real-time or previously captured by the OSI device of <figref idref="f0014">Figure 14</figref>;</li><li><figref idref="f0052">Figure 51</figref> illustrates a tear film image database for a patient;</li><li><figref idref="f0053">Figure 52</figref> illustrates a view images GUI screen showing an overlaid image of interference interactions of the interference signals from specularly reflected light from a patient's tear film overtop an image of the patient's eye for both the patient's left and right eyes side by side; and</li><li><figref idref="f0054">Figure 53</figref> illustrates the GUI screen of <figref idref="f0053">Figure 52</figref> with the images of the patient's eye toggled to show only the interference interactions of the interference signals from specularly reflected light from a patient's tear film.</li></ul>

<u>Detailed Description</u>

0021The embodiments set forth below represent the necessary information to enable those skilled in the art to practice the invention and illustrate the best mode of practicing the invention. Upon reading the following description in light of the accompanying drawing figures, those skilled in the art will understand the concepts of the invention and will recognize applications of these concepts not particularly addressed herein. It should be understood that these concepts and applications fall within the scope of the disclosure and the accompanying claims.

0022Embodiments of the detailed description include ocular surface interferometry (OSI) devices, systems, and methods for measuring a tear film layer thickness (TFLT) in a patient's ocular tear film. The OSI devices, systems, and methods can be used to measure the thickness of the lipid layer component (LLT) and/or the aqueous layer component (ALT) of the ocular tear film. "TFLT" as used herein includes LLT, ALT, or both LLT and ALT. "Measuring TFLT" as used herein includes measuring LLT, ALT, or both LLT and ALT. Measuring TFLT can be used in the diagnosis of a patient's tear film, including but not limited to lipid layer and aqueous layer deficiencies. These characteristics may be the cause or contributing factor to a patient experiencing dry eye syndrome (DES).

0023In this regard, embodiments disclosed herein include a light source that is controlled to direct light in the visible region to an ocular tear film. For example, the light source may be a Lambertian emitter that provides a uniform or substantially uniform intensity in all directions of emission. The light source is arranged such that light rays emitted from the light source are specularly reflected toward an imaging device from the tear film and undergo constructive and destructive interference interactions in the ocular tear film. An imaging device having a detection spectrum that includes the spectrum of the light source is focused on an area(s) of interest on the lipid layer of the tear film. The imaging device captures a first image of the interference interactions (i.e., modulation) of specularly reflected light rays from the illuminated tear film coming together by the focusing action of the imaging device. The imaging device then captures the interference signals representing the interference interactions of specularly reflected light from the tear film. The imaging device produces an output signal(s) representative of the interference signal in a first image. The first image may contain an interference signal for a given imaged pixel or pixels of the lipid layer by the imaging device. The output signal(s) can be processed and analyzed to measure a TFLT in the area or region of interest of the ocular tear film.

0024In this regard, <figref idref="f0003 f0004 f0005 f0006 f0007 f0008 f0009">Figures 4A-9</figref> illustrate a general embodiment of an ocular surface interferometry (OSI) device 30. Other embodiments will be described later in this application. In general, the OSI device 30 is configured to illuminate a patient's ocular tear film, capture images of interference interactions of specularly reflected light from the ocular tear film, and process and analyze the interference interactions to measure TFLT. As shown in <figref idref="f0003">Figure 4A</figref>, the exemplary OSI device 30 positioned in front of one of the patient's eye 32 is shown from a side view. A top view of the patient 34 in front of the OSI device 30 is illustrated in <figref idref="f0004">Figure 4B</figref>. The ocular tear film of a patient's eyes 32 is illuminated with a light source 36 (also referred to herein as "illuminator 36") and comprises a large area light source having a spectrum in the visible region adequate for TLFT measurement and correlation to dry eye. The illuminator 36 can be a white or multi-wavelength light source.

0025In this embodiment, the illuminator 36 is a Lambertian emitter and is adapted to be positioned in front of the eye 32 on a stand 38. As employed herein, the terms "Lambertian surface" and "Lambertian emitter" are defined to be a light emitter having equal or substantially equal (also referred to as "uniform" or substantially uniform) intensity in all directions. This allows the imaging of a uniformly or substantially uniformly bright tear film region for TFLT, as discussed in more detail in this disclosure. The illuminator 36 comprises a large surface area emitter, arranged such that rays emitted from the emitter are specularly reflected from the ocular tear film and undergo constructive and destructive interference in tear film layers therein. An image of the patient's 34 lipid layer is the backdrop over which the interference image is seen and it should be as spatially uniform as possible.

0026An imaging device 40 is included in the OSI device 30 and is employed to capture interference interactions of specularly reflected light from the patient's 34 ocular tear film when illuminated by the illuminator 36. The imaging device 40 may be a still or video camera, or other device that captures images and produces an output signal representing information in captured images. The output signal may be a digital representation of the captured images. The geometry of the illuminator 36 can be understood by starting from an imaging lens 42 of the imaging device 40 and proceeding forward to the eye 32 and then to the illuminator 36. The fundamental equation for tracing ray lines is Snell's law, which provides: <maths id="math0001" num=""><math display="block"><mi mathvariant="normal">n</mi><mn>1</mn><mspace width="1ex" /><mi>Sin</mi><mspace width="1ex" /><msub><mi mathvariant="normal">Θ</mi><mn>1</mn></msub><mo>=</mo><mi mathvariant="normal">n</mi><mn>2</mn><mspace width="1ex" /><mi>Sin</mi><mspace width="1ex" /><msub><mi mathvariant="normal">Θ</mi><mn>2</mn></msub><mo>,</mo></math><img file="EP2413699B1_D0001.tif" /></maths> where "n1" and "n2" are the indexes of refraction of two mediums containing the ray, and Θ₁ and Θ₂ is the angle of the ray relative to the normal from the transition surface. As illustrated in <figref idref="f0005">Figure 5</figref>, light rays 44 are directed by the illuminator 36 to an ocular tear film 46. In the case of specularly reflected light 48 that does not enter a lipid layer 50 and instead reflects from an anterior surface 52 of the lipid layer 50, Snell's law reduces down to Θ₁ = Θ₂, since the index of refraction does not change (i.e., air in both instances). Under these conditions, Snell's law reduces to the classical law of reflection such that the angle of incidence is equal and opposite to the angle of reflectance.

0027Some of the light rays 54 pass through the anterior surface 52 of the lipid layer 50 and enter into the lipid layer 50, as illustrated in <figref idref="f0005">Figure 5</figref>. As a result, the angle of these light rays 54 (i.e., Θ₃) normal to the anterior surface 52 of the lipid layer 50 will be different than the angle of the light rays 44 (Θ₁) according to Snell's law. This is because the index of refraction of the lipid layer 50 is different than the index of refraction of air. Some of the light rays 54 passing through the lipid layer 50 will specularly reflect from the lipid layer-to-aqueous layer transition 56 thereby producing specularly reflected light rays 58. The specularly reflected light rays 48, 58 undergo constructive and destructive interference anterior of the lipid layer 50. The modulations of the interference of the specularly reflected light rays 48, 58 superimposed on the anterior surface 52 of the lipid layer 50 are collected by the imaging device 40 when focused on the anterior surface 52 of the lipid layer 50. Focusing the imaging device 40 on the anterior surface 52 of the lipid layer 50 allows capturing of the modulated interference information at the plane of the anterior surface 52. In this manner, the captured interference information and the resulting calculated TFLT from the interference information is spatially registered to a particular area of the tear film 46 since that the calculated TFLT can be associated with such particular area, if desired.

0028The thickness of the lipid layer 50 ('d₁') is a function of the interference interactions between specularly reflected light rays 48, 58. The thickness of the lipid layer 50 ('d₁') is on the scale of the temporal (or longitudinal) coherence of the light source 30. Therefore, thin lipid layer films on the scale of one wavelength of visible light emitted by the light source 30 offer detectable colors from the interference of specularly reflected light when viewed by a camera or human eye. The colors may be detectable as a result of calculations performed on the interference signal and represented as a digital values including but not limited to a red-green-blue (RGB) value in the RGB color space. Quantification of the interference of the specularly reflected light can be used to measure LLT. The thicknesses of an aqueous layer 60 ('d₂') can also be determined using the same principle. Some of the light rays 54 (not shown) passing through the lipid layer 50 can also pass through the lipid-to-aqueous layer transition 56 and enter into the aqueous layer 60 specularly reflecting from the aqueous-to-mucin/cornea layer transition 62. These specular reflections also undergo interference with the specularly reflected light rays 48, 58. The magnitude of the reflections from each interface depends on the refractive indices of the materials as well as the angle of incidence, according to Fresnel's equations, and so the depth of the modulation of the interference interactions is dependent on these parameters, thus so is the resulting color.

0029Turning back to <figref idref="f0003">Figures 4A</figref> and <figref idref="f0004">4B</figref>, the illuminator 36 in this embodiment is a broad spectrum light source covering the visible region between about 400 nm to about 700 nm. The illuminator 36 contains an arced or curved housing 64 (see <figref idref="f0004">Figure 4B</figref>) into which individual light emitters are mounted, subtending an arc of approximately 130 degrees from the optical axis of the eye 32 (see <figref idref="f0004">Figure 4B</figref>). A curved surface may present better uniformity and be more efficient, as the geometry yields a smaller device to generating a given intensity of light. The total power radiated from the illuminator 36 should be kept to a minimum to prevent accelerated tear evaporation. Light entering the pupil can cause reflex tearing, squinting, and other visual discomforts, all of which affect TFLT measurement accuracy.

0030In order to prevent alteration of the proprioceptive senses and reduce heating of the tear film 46, incident power and intensity on the eye 32 may be minimized and thus, the step of collecting and focusing the specularly reflected light may carried out by the imaging device 40. The imaging device 40 may be a video camera, slit lamp microscope, or other observation apparatus mounted on the stand 38, as illustrated in <figref idref="f0003">Figures 4A</figref> and <figref idref="f0004">4B</figref>. Detailed visualization of the image patterns of the tear film 46 involves collecting the specularly reflected light 66 and focusing the specularly reflected light at the lipid layer 52 such that the interference interactions of the specularly reflected light from the ocular tear film are observable.

0031Against the backdrop of the OSI device 30 in <figref idref="f0003">Figures 4A</figref> and <figref idref="f0004">4B</figref>, <figref idref="f0006">Figure 6</figref> illustrates a flowchart discussing how the OSI device 30 can be used to obtain interference interactions of specularly reflected light from the tear film 46, which can be used to measure TFLT. Interference interactions of specularly reflected light from the tear film 46 are first obtained and discussed before measurement of TFLT is discussed. In this embodiment as illustrated in <figref idref="f0006">Figure 6</figref>, the process starts by adjusting the patient 32 with regard to an illuminator 36 and an imaging device 40 (block 70). The illuminator 36 is controlled to illuminate the patient's 34 tear film 46. The imaging device 40 is controlled to be focused on the anterior surface 52 of the lipid layer 50 such that the interference interactions of specularly reflected light from the tear film 46 are collected and are observable. Thereafter, the patient's 34 tear film 46 is illuminated by the illuminator 36 (block 72).

0032The imaging device 40 is then controlled and focused on the lipid layer 50 to collect specularly reflected light from an area or region of interest on a tear film as a result of illuminating the tear film with the illuminator 36 in a first image (block 74, <figref idref="f0006">Figure 6</figref>). An example of the first image by the illuminator 36 is provided in <figref idref="f0007">Figure 7</figref>. As illustrated therein, a first image 79 of a patient's eye 80 is shown that has been illuminated with the illuminator 36. The illuminator 36 and the imaging device 40 may be controlled to illuminate an area or region of interest 81 on a tear film 82 that does not include a pupil 83 of the eye 80 so as to reduce reflex tearing. Reflex tearing will temporarily lead to thicker aqueous and lipid layers, thus temporarily altering the interference signals of specularly reflected light from the tear film 82. As shown in <figref idref="f0007">Figure 7</figref>, when the imaging device 40 is focused on an anterior surface 86 of the lipid layer 88 of the tear film 82, interference interactions 85 of the interference signal of the specularly reflected light from the tear film 82 as a result of illumination by the illuminator 36 are captured in the area or region of interest 81 in the first image 79. The interference interactions 85 appear to a human observer as colored patterns as a result of the wavelengths present in the interference of the specularly reflected light from the tear film 82.

0033However, the background signal is also captured in the first image 79. The background signal is added to the specularly reflected light in the area or region of interest 81 and included outside the area or region of interest 81 as well. Background signal is light that is not specularly reflected from the tear film 82 and thus contains no interference information. Background signal can include stray and ambient light entering into the imaging device 40, scattered light from the patient's 34 face, eyelids, and/or eye 80 structures outside and beneath the tear film 82 as a result of stray light, ambient light and diffuse illumination by the illuminator 36, and images of structures beneath the tear film 82. For example, the first image 79 includes the iris of the eye 80 beneath the tear film 82. Background signal adds a bias (i.e., offset) error to the captured interference of specularly reflected light from the tear film 82 thereby reducing its signal strength and contrast. Further, if the background signal has a color hue different from the light of the light source, a color shift can also occur to the interference of specularly reflected light from the tear film 82 in the first image 79. The imaging device 40 produces a first output signal that represents the light rays captured in the first image 79. Because the first image 79 contains light rays from specularly reflected light as well as the background signal, the first output signal produced by the imaging device 40 from the first image 79 will contain an interference signal representing the captured interference of the specularly reflected light from the tear film 82 with a bias (i.e., offset) error caused by the background signal. As a result, the first output signal analyzed to measure TFLT may contain error as a result of the background signal bias (i.e., offset) error.

0034Thus, in this embodiment, the first output signal generated by the imaging device 40 as a result of the first image 79 is processed to subtract or substantially subtract the background signal from the interference signal to reduce error before being analyzed to measure TFLT. This is also referred to as "background subtraction." Background subtraction is the process of removing unwanted reflections from images. In this regard, the imaging device 40 is controlled to capture a second image 90 of the tear film 82 when not illuminated by the illuminator 36, as illustrated by example in <figref idref="f0008">Figure 8</figref>. The second image 90 should be captured using the same imaging device 40 settings and focal point as when capturing the first image 79 so that the first image 79 and second images 90 forms corresponding image pairs captured within a short time of each other. The imaging device 40 produces a second output signal containing background signal present in the first image 79 (block 76 in <figref idref="f0006">Figure 6</figref>). To eliminate or reduce this background signal from the first output signal, the second output signal is subtracted from the first output signal to produce a resulting signal (block 77 in <figref idref="f0006">Figure 6</figref>). The image representing the resulting signal in this example is illustrated in <figref idref="f0009">Figure 9</figref> as resulting image 92. Thus, in this example, background subtraction involves two images 79, 90 to provide a frame pair where the two images 79, 90 are subtracted from each other, whereby specular reflection from the tear film 82 is retained, and while diffuse reflections from the iris and other areas are removed in whole or part.

0035As illustrated in <figref idref="f0009">Figure 9</figref>, the resulting image 92 contains an image of the isolated interference 94 of the specularly reflected light from the tear film 82 with the background signal eliminated or reduced (block 78 in <figref idref="f0006">Figure 6</figref>). In this manner, the resulting signal (representing the resulting image 92 in <figref idref="f0009">Figure 9</figref>) includes an interference signal having signal improved purity and contrast in the area or region of interest 81 on the tear film 82. As will be discussed later in this application, the resulting signal provides for accurate analysis of interference interactions from the interference signal of specular reflections from the tear film 82 to in turn accurately measure TFLT. Any method or device to obtain the first and second images of the tear film 82 and perform the subtraction of background signal in the second image 90 from the first image 79 may be employed. Other specific examples are discussed throughout the remainder of this application.

0036An optional registration function may be performed between the first image(s) 79 and the second image(s) 90 before subtraction is performed to ensure that an area or point in the second image(s) 90 to be subtracted from the first image(s) 79 is for an equivalent or corresponding area or point on the first image(s) 79. For example, a set of homologous points may be taken from the first and second images 79, 90 to calculate a rigid transformation matrix between the two images. The transformation matrix allows one point on one image (e.g., x1, y1) to be transformed to an equivalent two-dimensional (2D) image on the other image (e.g., x2, y2). For example, the Matlab® function "cp2tform" can be employed in this regard. Once the transformation matrix is determined, the transformation matrix can be applied to every point in the first and second images, and then each re-interpolated at the original points. For example, the Matlab® function "imtransform" can be employed in this regard. This allows a point from the second image (e.g., x2, y2) to be subtracted from the correct, equivalent point (e.g., x1, y1) on the first image(s) 79, in the event there is any movement in orientation or the patient's eye between the capture of the first and second images 79, 90. The first and second images 79, 90 should be captured close in time.

0037Note that while this example discusses a first image and a second image captured by the imaging device 40 and a resulting first output signal and second output signal, the first image and the second image may comprise a plurality of images taken in a time-sequenced fashion. If the imaging device 40 is a video camera, the first and second images may contain a number of sequentially-timed frames governed by the frame rate of the imaging device 40. The imaging device 40 produces a series of first output signals and second output signals. If more than one image is captured, the subtraction performed in a first image should ideally be from a second image taken immediately after the first image so that the same or substantially the same lighting conditions exist between the images so the background signal in the second image is present in the first image. The subtraction of the second output signal from the first output signal can be performed in real time. Alternatively, the first and second output signals can be recorded and processed at a later time. The illuminator 36 may be controlled to oscillate off and on quickly so that first and second images can be taken and the second output signal subtraction from the first output signal be performed in less than one second. For example, if the illuminator 36 oscillates between on and off at 30 Hz, the imaging device 40 can be synchronized to capture images of the tear film 46 at 60 frames per second (fps). In this regard, thirty (30) first images and thirty (30) second images can be obtained in one second, with each pair of first and second images taken sequentially.

0038After the interference of the specularly reflected light is captured and a resulting signal containing the interference signal is produced and processed, the interference signal or representations thereof can be compared against a tear film layer interference model to measure TFLT. The interference signal can be processed and converted by the imaging device into digital red-green-blue (RGB) component values which can be compared to RGB component values in a tear film interference model to measure tear film TFLT. The tear film interference model is based on modeling the lipid layer of the tear film in various LLTs and representing resulting interference interactions in the interference signal of specularly reflected light from the tear film model when illuminated by the light source. The tear film interference model can be a theoretical tear film interference model where the particular light source, the particular imaging device, and the tear film layers are modeled mathematically, and the resulting interference signals for the various LLTs recorded when the modeled light source illuminates the modeled tear film layers recorded using the modeled imaging device. The settings for the mathematically modeled light source and imaging device should be replicated in the illuminator 36 and imaging device 40 used in the OSI device 30. Alternatively, the tear film interference model can be based on a phantom tear film model, comprised of physical phantom tear film layers wherein the actual light source is used to illuminate the phantom tear film model and interference interactions in the interference signal representing interference of specularly reflected light are empirically observed and recorded using the actual imaging device.

0039The aqueous layer may be modeled in the tear film interference model to be of an infinite, minimum, or varying thickness. If the aqueous layer is modeled to be of an infinite thickness, the tear film interference model assumes no specular reflections occur from the aqueous-to-mucin layer transition 62 (see <figref idref="f0005">Figure 5</figref>). If the aqueous layer 62 is modeled to be of a certain minimum thickness (e.g., ≥ 2 µm), the specular reflection from the aqueous-to-mucin layer transition 62 may be considered negligible on the effect of the convolved RGB signals produced by the interference signal. In either case, the tear film interference model will only assume and include specular reflections from the lipid-to-aqueous layer transition 56. Thus, these tear film interference model embodiments allow measurement of LLT regardless of ALT. The interference interactions in the interference signal are compared to the interference interactions in the tear film interference model to measure LLT.

0040Alternatively, if the aqueous layer 60 is modeled to be of varying thicknesses, the tear film interference model additionally includes specular reflections from the aqueous-to-mucin layer transition 62 in the interference interactions. As a result, the tear film interference model will include two-dimensions of data comprised of interference interactions corresponding to various LLT and ALT combinations. The interference interactions from the interference signal can be compared to interference interactions in the tear film interference model to measure both LLT and ALT. More information regarding specific tear film interference models will be described later in this application.

0041In the above described embodiment in <figref idref="f0006 f0007 f0008 f0009">Figures 6-9</figref>, the second image 90 of the tear film 82 containing background signal is captured when not illuminated by the illuminator 36. Only ambient light illuminates the tear film 82 and eye 80 structures beneath. Thus, the second image 90 and the resulting second output signal produced by the imaging device 40 from the second image 90 does not include background signal resulting from scattered light from the patient's face and eye structures as a result of diffuse illumination by the illuminator 36. Only scattered light resulting from ambient light is included in the second image 90. However, scattered light resulting from diffuse illumination by the illuminator 36 is included in background signal in the first image 79 containing the interference interactions of specularly reflected light from the tear film 82. Further, because the first image 79 is captured when the illuminator 36 is illuminating the tear film, the intensity of the eye structures beneath the tear film 82 captured in the first image 79, including the iris, are brighter than captured in the second image 90. Thus, in other embodiments described herein, the imaging device 40 is controlled to capture a second image of the tear film 82 when obliquely illuminated by the illuminator 36. As a result, the captured second image additionally includes background signal from scattered light as a result of diffuse illumination by the illuminator 36 as well as a higher intensity signal of the eye directly illuminated structures beneath the tear film 82. Thus, when the second output signal is subtracted from the first output signal, the higher intensity eye structure background and the component of background signal representing scattered light as a result of diffuse illumination by the illuminator 36, as well as ambient and stray light, are subtracted or substantially subtracted from the resulting signal thereby further increasing the interference signal purity and contrast in the resulting signal. The resulting signal can then be processed and analyzed to measure TFLT, as will be described in detail later in this application.

0042In this regard, <figref idref="f0010 f0011 f0012">Figures 10-12</figref> illustrate an embodiment for illuminating and capturing interference of specularly reflected light from the tear film. In this embodiment, the second image is captured when the tear film is obliquely illuminated by the illuminator 36 using illumination that possesses the same or nearly the same average geometry and illuminance level as used to produce specularly reflected light from a tear film. In this manner, the background signal captured in the second image contains the equivalent background signal present in the first image including scattered light from the tear film and patient's eye as a result of diffuse illumination by the illuminator 36. The second image also includes a representative signal of eye structure beneath the tear film because of the equivalent lighting when the illuminator 36 is activated when capturing the second image. In this embodiment, a "tiled" or "tiling" illumination of the tear film is provided. Tiling allows a light source to illuminate a sub-area(s) of interest on the tear film to obtain specularly reflected light while at the same time diffusely illuminating adjacent sub-area(s) of interest of the tear film to obtain scattered light as a result of diffuse illumination by the illuminator 36. In this manner, the subtracted background signal includes scattered light as a result of diffuse illumination by the illuminator 36 to allow further reduction of offset bias (i.e., offset) error and to thereby increase interference signal purity and contrast.

0043In this regard, as illustrated in <figref idref="f0010">Figure 10</figref>, the process starts by adjusting the patient 34 with regard to the illuminator 36 and the imaging device 40 (block 100). The illuminator 36 is controlled to illuminate the patient's 34 tear film. The imaging device 40 is located appropriately and is controlled to be focused on the lipid layer such that the interference interactions of specularly reflected light from the tear film are observable when the tear film is illuminated. Thereafter, the lighting pattern of the illuminator 36 is controlled in a first "tiling" mode to produce specularly reflected light from a first area(s) of interest of the tear film while diffusely illuminating an adjacent, second area(s) of interest of the tear film (block 102). As will be discussed in more detail later in this application, the illuminator 36 may be controlled to turn on only certain lighting components in the illuminator 36 to control the lighting pattern.

0044An example of a first image 120 captured of a patient's eye 121 and tear film 123 by the imaging device 40 when the illuminator 36 produces a light pattern in the first mode is illustrated by example in <figref idref="f0011">Figure 11A</figref>. In this example, the illuminator 36 is controlled to provide a first tiled illumination pattern in an area or region of interest 122 on the tear film 123. While illumination of the tear film 123 in the first mode, the imaging device 40 captures the first image 120 of the patient's eye 121 and the tear film 123 (block 104). As illustrated in <figref idref="f0011">Figure 11A</figref>, the first image 120 of the patient's eye 121 has been illuminated so that specularly reflected light is produced in first portions 126A in the area or region of interest 122 of the tear film 123. The interference signal(s) from the first portions 126A include interference from specularly reflected light along with additive background signal, which includes scattered light signal as a result of diffuse illumination from the illuminator 36. Again, the illuminator 36 and the imaging device 140 may be controlled to illuminate the tear film 123 that does not include the pupil of the eye 121 so as to reduce reflex tearing. The illuminator 36 may be flashed in block 102 to produce specularly reflected light from the first portions 126A, whereby the imaging device 40 is synchronized with the flashing of the illuminator 36 in block 104 to capture the first image 120 of the patient's eye 121 and the tear film 123.

0045Also during the first mode, the illuminator 36 light pattern obliquely illuminates second, adjacent second portions 128A to the first portions 126A in the area or region of interest 122, as shown in the first image 120 in <figref idref="f0011">Figure 11A</figref>. The second portions 128A include comparable background offset present in the first portion(s) 126A, which includes scattered light signal as a result of diffuse illumination from the illuminator 36 since the illuminator 36 is turned on when the first image 120 is captured by the imaging device 40. Further, the eye 121 structures beneath the tear film 123 are captured in the second portions 128A due to the diffuse illumination by the illuminator 36. This is opposed to the second image 90 of <figref idref="f0009">Figure 9</figref>, where diffuse illumination by the illuminator 36 is not provided to the tear film when the second image 90 is obtained. Thus, in this embodiment, the area or region of interest 122 of the tear film 123 is broken into two portions at the same time: first portions 126A producing specularly reflected light combined with background signal, and second portions 128A diffusedly illuminated by the illuminator 36 and containing background signal, which includes scattered light from the illuminator 36. The imaging device 40 produces a first output signal that contains a representation of the first portions 126A and the second portions 128A.

0046Next, the illuminator 36 is controlled in a second mode to reverse the lighting pattern from the first mode when illuminating the tear film 123 (block 106, <figref idref="f0010">Figure 10</figref>). A second image 130 is captured of the tear film 121 is captured in the second mode of illumination, as illustrated by example in <figref idref="f0011">Figure 11B</figref> (block 108, <figref idref="f0010">Figure 10</figref>). As shown in the second image 130 in <figref idref="f0011">Figure 11B</figref>, the second portions 128A in the first image 120 of <figref idref="f0011">Figure 11A</figref> are now second portions 128B in the second image 130 in <figref idref="f0011">Figure 11B</figref> containing specularly reflected light from the tear film 123 with additive background signal. The first portions 126A in the first image 120 of <figref idref="f0011">Figure 11A</figref> are now first portions 126B in the second image 130 in <figref idref="f0011">Figure 11B</figref> containing background signal without specularly reflected light. Again, the background signal in the first portions 126B includes scattered light signal as a result of diffuse illumination by the illuminator 36. The imaging device 40 produces a second output signal of the second image 130 in <figref idref="f0011">Figure 11B</figref>. The illuminator 36 may also be flashed in block 106 to produce specularly reflected light from the second portions 128B, whereby the imaging device 40 is synchronized with the flashing of the illuminator 36 in block 106 to capture the second image 130 of the patient's eye 121 and the tear film 123.

0047The first and second output signals can then be combined to produce a resulting signal comprised of the interference signal of the specularly reflected light from the tear film 123 with background signal subtracted or substantially removed from the interference signal (block 110, <figref idref="f0010">Figure 10</figref>). A resulting image is produced as a result having interference information from the specularly reflected light from the area or region of interest 122 of the tear film 123 with background signal eliminated or reduced, including background signal resulting from scattered light from diffuse illumination by the illuminator 36 (block 112, <figref idref="f0010">Figure 10</figref>). An example of a resulting image 132 in this regard is illustrated in <figref idref="f0012">Figure 12</figref>. The resulting image 132 represents the first output signal represented by the first image 120 in <figref idref="f0011">Figure 11A</figref> combined with the second output signal represented by the second image 130 in <figref idref="f0011">Figure 11B</figref>. As illustrated in <figref idref="f0012">Figure 12</figref>, interference signals of specularly reflected light from the tear film 123 are provided for both the first and second portions 126, 128 in the area or region of interest 122. The background signal has been eliminated or reduced. As can be seen in <figref idref="f0012">Figure 12</figref>, the signal purity and contrast of the interference signal representing the specularly reflected light from the tear film 123 from first and second portions 126, 128 appears more vivid and higher in contrast than the interference interaction 94 in <figref idref="f0009">Figure 9</figref>, for example.

0048In the discussion of the example first and second images 120, 130 in <figref idref="f0011">Figures 11A and 11B</figref> above, each first portion 126 can be thought of as a first image, and each second portion 128 can be thought of as a second image. Thus, when the first and second portions 126A, 128B are combined with corresponding first and second portions 126B, 128A, this is akin to subtracting second portions 126B, 128A from the first portions 12A, 128B, respectively.

0049In the example of <figref idref="f0010 f0011 f0012">Figures 10-12</figref>, the first image and second images 120, 130 contain a plurality of portions or tiles. The number of tiles depends on the resolution of lighting interactions provided for and selected for the illuminator 36 to produce the first and second modes of illumination to the tear film 123. The illumination modes can go from one extreme of one tile to any number of tiles desired. Each tile can be the size of one pixel in the imaging device 40 or areas covering more than one pixel depending on the capability of the illuminator 36 and the imaging device 40. The number of tiles can affect accuracy of the interference signals representing the specularly reflected light from the tear film. Providing too few tiles in a tile pattern can limit the representative accuracy of the average illumination geometry that produces the scattered light signal captured by the imaging device 40 in the portions 128A and 126B for precise subtraction from portions 128B and 126A respectively.

0050Note that while this example in <figref idref="f0010 f0011 f0012">Figures 10-12</figref> discusses a first image and a second image captured by the imaging device 40 and a resulting first output signal and second output signal, the first image and the second image may comprise a plurality of images taken in a time-sequenced fashion. If the imaging device 40 is a video camera, the first and second images may contain a number of sequentially-timed frames governed by the frame rate of the imaging device 40. The imaging device 40 produces a series of first output signals and second output signals. If more than one image is captured, the subtraction performed in a first image should ideally be from a second image taken immediately after the first image so that the same or substantially the same lighting conditions exist between the images so the background signal in the second image is present in the first image, and more importantly, so that movement of the eye and especially of the tear-film dynamic is minimal between subtracted frames. The subtraction of the second output signal from the first output signal can be performed in real time. Alternatively, the first and second output signals can be recorded and processed at a later time.

0051Other optical tiling patterns are possible other than the "teeth" style tiling pattern illustrated in <figref idref="f0011 f0012">Figures 11A-12</figref>. <figref idref="f0013">Figures 13A and 13B</figref> illustrate an alternative tiling mode embodiment via illustrations of images of an eye 140 and tear film 142. In this embodiment, a concentric optical tiling pattern is provided by the illuminator 36 for illuminating the tear film 142. The interference interactions of the specularly reflected light from the tear film 142 are captured by the imaging device 40. As illustrated in <figref idref="f0013">Figure 13A</figref>, a first image 144 is taken of an area or region of interest 146 on the tear film 142 during a first mode of the illuminator 36. The illuminator 36 is controlled to produce a first lighting pattern in the first mode such that a center portion 148 of the area or region of interest 146 of the tear film 142 produces specularly reflected light from the tear film 142. The center portion 148 includes specularly reflected light from the tear film 142 along with background signal, including scattered light signal from diffuse illumination of the tear film 142 by the illuminator 36. Background signal is produced from the edge portions 152 of the area or region of interest 146. The imaging device 140 produces a first output signal representative of the first image 144 in <figref idref="f0013">Figure 13A</figref>.

0052In a second mode of the illuminator 36, as illustrated by the representative second image 160 in <figref idref="f0013">Figure 13B</figref>, the illuminator 36 is controlled to reverse the lighting pattern for illuminating the tear film 142 from the first mode. Specularly reflected light is now produced from the edge portions 152 in the area or region of interest 146, which includes additive background signal. The center portion 148 now produces only background signal. In this manner, the center portion 148 and the edge portions 152 are concentric portions. The imaging device 40 produces a second output signal representative of the second image 160 in <figref idref="f0013">Figure 13B</figref>.

0053The first and second output signals can then be combined to produce a resulting signal comprised of the interference signal of the specularly reflected light from the tear film 142 for the entire area or region of interest 146 with background signal subtracted or substantially removed from the interference signal. A resulting image (not shown) similar to <figref idref="f0012">Figure 12</figref> can be produced as a result of having interference information from the specularly reflected light from the area or region of interest 146 from the tear film 142 with background signal eliminated or reduced, including background signal resulting from scattered light from diffuse illumination by the illuminator 36. The resulting image can then be processed and analyzed to measure TFLT. In the example of <figref idref="f0013">Figures 13A and 13B</figref>, the illuminator 36 is controlled in the first and second modes such that the relationship of the areas between the center portion 148 and the edge portion 152 is balanced to be approximately 50%/50% so that an equal balance of diffuse illumination from the illuminator 36 is provided in both modes to portions of the tear film 142 that do not produce specularly reflected light. However, other balance percentages can be employed.

0054Alternatively, a small-scale scanning of the ocular tear film can be employed to obtain interference of specularly reflected light from the tear film to obtain a high signal strength and contrast of an interference signal without providing tiled illumination patterns or diffuse light from the illuminator 36. For example, the area or region of interest imaged on the ocular tear film could be made very small down to the lowest resolution of the imaging device 40 (e.g., one pixel). In this manner, virtually no diffuse illumination is provided from the illuminator 36 to the area or region of interest on the patient's tear film when illuminated. Background signal captured in the image of the specularly reflected light from the tear film would be negligible compared to the level of specularly reflected light captured in the image. Thus, no subtraction of multiple images may need to be performed. The illuminator 36 would be controlled to scan the desired portions of the tear film for sequential image capture, with each scan capturing an image of specularly reflected light from a small area or region of interest. Each scanned image can then be assembled to produce an overall image of specularly reflected light from the tear film with negligible background signal and processed and analyzed to measure TFLT.

<u>EXEMPLARY OSI DEVICE</u>

0055The above discussed illustrations provide examples of illuminating and imaging a patient's TFLT. These principles are described in more detail with respect to a specific example of an OSI device 170 illustrated in <figref idref="f0014 f0015 f0016 f0017 f0018 f0019 f0020 f0021 f0022 f0023 f0024 f0025 f0026 f0027 f0028 f0029 f0030 f0031 f0032 f0033 f0034 f0035 f0036 f0037 f0038 f0039 f0040 f0041 f0042 f0043 f0044 f0045 f0046 f0047 f0048 f0049 f0050 f0051">Figures 14-50</figref> and described below throughout the remainder of this application. The OSI device 170 can illuminate a patient's tear film, capture interference information from the patient's tear film, and process and analyze the interference information to measure TFLT. Further, the OSI device 170 includes a number of optional pre-processing features that may be employed to process the interference signal in the resulting signal to enhance TFLT measurement. The OSI device 170 may include a display and user interface to allow a physician or technician to control the OSI device 170 to image a patient's eye and tear film and measure the patient's TFLT.

<u>Illumination and Imaging</u>

0056In this regard, <figref idref="f0014">Figure 14</figref> illustrates a perspective view of the OSI device 170. The OSI device 170 is designed to facilitate imaging of the patient's ocular tear film and processing and analyzing the images to determine characteristics regarding a patient's tear film. The OSI device 170 includes an imaging device and light source in this regard, as will be described in more detail below. As illustrated in <figref idref="f0014">Figure 14</figref>, the OSI device 170 is comprised generally of a housing 172, a display monitor ("display") 174, and a patient head support 176. The housing 172 may be designed for table top placement. The housing 172 rests on a base 178 in a fixed relationship. As will be discussed in more detail below, the housing 172 houses an imaging device and other electronics, hardware, and software to allow a clinician to image a patient's ocular tear film. A light source 173 (also referred to herein as "illuminator 173") is also provided in the housing 172 and provided behind a diffusing translucent window 175. The translucent window 175 may be a flexible, white, translucent acrylic plastic sheet.

0057To image a patient's ocular tear film, the patient places his or her head in the patient head support 176 and rests his or her chin on a chin rest 180. The chin rest 180 can be adjusted to align the patient's eye and tear film with the imaging device inside the housing 172, as will be discussed in more detail below. The chin rest 180 may be designed to support up to two (2) pounds of weight, but such is not a limiting factor. A transparent window 177 allows the imaging device inside the housing 172 to have a clear line of sight to a patient's eye and tear film when the patient's head is placed in the patient head support 176. The OSI device 170 is designed to image one eye at a time, but can be configured to image both eyes of a patient, if desired.

0058In general, the display 174 provides input and output from the OSI device 170. For example, a user interface can be provided on the display 174 for the clinician to operate the OSI device 170 and to interact with a control system provided in the housing 172 that controls the operation of the OSI device 170, including an imaging device, an imaging device positioning system, a light source, other supporting hardware and software, and other components. For example, the user interface can allow control of imaging positioning, focus of the imaging device, and other settings of the imaging device for capturing images of a patient's ocular tear film. The control system may include a general purpose microprocessor or computer with memory for storage of data, including images of the patient's eye and tear film. The microprocessor should be selected to provide sufficient processing speed to process images of the patient's tear film and generate output characteristic information about the tear film (e.g., one minute per twenty second image acquisition). The control system may control synchronization of activation of the light source and the imaging device to capture images of areas of interest on the patient's ocular tear film when properly illuminated. Various input and output ports and other devices can be provided, including but not limited to a joystick for control of the imaging device, USB ports, wired and wireless communication including Ethernet communication, a keyboard, a mouse, speaker(s), etc. A power supply is provided inside the housing 172 to provide power to the components therein requiring power. A cooling system, such as a fan, may also be provided to cool the OSI device 170 from heat generating components therein.

0059The display 174 is driven by the control system to provide information regarding a patient's imaged tear film, including TFLT. The display 174 also provides a graphical user interface (GUI) to allow a clinician or other user to control the OSI device 170. To allow for human diagnosis of the patient's tear film, images of the patient's ocular tear film taken by the imaging device in the housing 172 can also be displayed on the display 174 for review by a clinician, as will be illustrated and described in more detail below. The images displayed on the display 174 may be real-time images being taken by the imaging device, or may be previously recorded images stored in memory. To allow for different orientations of the OSI device 170 to provide a universal configuration for manufacturing, the display 174 can be rotated about the base 178. The display 174 is attached to a monitor arm 182 that is rotatable about the base 178, as illustrated. The display 174 can be placed opposite of the patient head support 176, as illustrated in <figref idref="f0014">Figure 14</figref>, if the clinician desires to sit directly across from the patient. Alternatively, display 174 can be rotated either left or right about the X-axis to be placed adjacent to the patient head support 176. The display 174 may be a touch screen monitor to allow a clinician or other user to provide input and control to the control system inside the housing 172 directly via touch of the display 174 for control of the OSI device 170. The display 174 illustrated in <figref idref="f0014">Figure 14</figref> is a fifteen inch (15") flat panel liquid crystal display (LCD). However, the display 174 may be provided of any type or size, including but not limited to a cathode ray tube (CRT), plasma, LED, OLED, projection system, etc.

0060<figref idref="f0015">Figure 15</figref> illustrates a side view of the OSI device 170 of <figref idref="f0014">Figure 14</figref> to further illustrate imaging of a patient's eye and ocular tear film. As illustrated therein, a patient places their head 184 in the patient head support 176. More particularly, the patient places their forehead 186 against a headrest 188 provided as part of the patient head support 176. The patient places their chin 190 in the chin rest 180. The patient head support 176 is designed to facilitate alignment of a patient's eye 192 with the OSI device 170, and in particular, an imaging device 194 (and illuminator) shown as being provided inside the housing 172. The chin rest 180 can be adjusted higher or lower to move the patient's eye 192 with respect to the OSI device 170.

0061As shown in <figref idref="f0016">Figure 16</figref>, the imaging device 194 is used to image the patient's ocular tear film to determine characteristics of the patient's tear film. In particular, the imaging device 194 is used to capture interference interactions of the specularly reflected light from the patient's tear film when illuminated by a light source 196 (also referred to herein as "illuminator 196") as well as background signal. As previously discussed, background signal may be captured when the illuminator 196 is illuminating or not illuminating a patient's tear film. In the OSI device 170, the imaging device 194 is the "The Imaging Source" model DFK21BU04 charge coupling device (CCD) digital video camera 198, but many types of metrological grade cameras or imaging devices can be provided. A CCD camera enjoys characteristics of efficient light gathering, linear behavior, cooled operation, and immediate image availability. A linear imaging device is one that provides an output signal representing a captured image which is precisely proportional to the input signal from the captured image. Thus, use of a linear imaging device (e.g., gamma correction set to 1.0, or no gamma correction) provides undistorted interference data which can then be analyzed using linear analysis models. In this manner, the resulting images of the tear film do not have to be linearized before analysis, thus saving processing time. Gamma correction can then be added to the captured linear images for human-perceptible display on a non-linear display 174 in the OSI device 170. Alternatively, the opposite scenario could be employed. That is, a non-linear imaging device or non-linear setting would be provided to capture tear film images, wherein the non-linear data representing the interference interactions of the interference signal can be provided to a non-linear display monitor without manipulation to display the tear film images to a clinician. The non-linear data would be linearized for tear film processing and analysis to estimate tear film layer thickness.

0062The video camera 198 is capable of producing lossless full motion video images of the patient's eye. As illustrated in <figref idref="f0016">Figure 16</figref>, the video camera 198 has a depth of field defined by the angle between rays 199 and the lens focal length that allows the patient's entire tear film to be in focus simultaneously. The video camera 198 has an external trigger support so that the video camera 198 can be controlled by a control system to image the patient's eye. The video camera 198 includes a lens that fits within the housing 172. The video camera 198 in this embodiment has a resolution of 640x480 pixels and is capable of frame rates up to sixty (60) frames per second (fps). The lens system employed in the video camera 198 images a 16 x 12 mm dimension in a sample plane onto an active area of a CCD detector within the video camera 198. As an example, the video camera 198 may be the DBK21AU04 Bayer VGA (640x480) video camera using a Pentax VS-LD25 Daitron 25-mm fixed focal length lens. Other camera models with alternate pixel size and number, alternate lenses, (etc) may also be employed.

0063Although a video camera 198 is provided in the OSI device 170, a still camera could also be used if the frame rate is sufficiently fast enough to produce high quality images of the patient's eye. High frame rate in frames per second (fps) facilitate high quality subtraction of background signal from a captured interference signal representing specularly reflected light from a patient's tear film, and may provide less temporal (i.e., motion) artifacts (e.g., motion blurring) in captured images, resulting in high quality captured images. This is especially the case since the patient's eye may move irregularly as well as blinking, obscuring the tear film from the imaging device during examination.

0064A camera positioning system 200 is also provided in the housing 172 of the OSI device 170 to position the video camera 198 for imaging of the patient's tear film. The camera positioning system 200 is under the control of a control system. In this manner, a clinician can manipulate the position of the video camera 198 to prepare the OSI device 170 to image the patient's tear film. The camera positioning system 200 allows a clinician and/or control system to move the video camera 198 between different patients' eyes 192, but can also be designed to limit the range of motion within designed tolerances. The camera positioning system 200 also allows for fine tuning of the video camera 198 position. The camera positioning system 200 includes a stand 202 attached to a base 204. A linear servo or actuator 206 is provided in the camera positioning system 200 and connected between the stand 202 and a camera platform 207 supporting the video camera 198 to allow the video camera 198 to be moved in the vertical (i.e., Y-axis) direction.

0065In this embodiment of the OSI device 170, the camera positioning system 200 may not allow the video camera 198 to be moved in the X-axis or the Z-axis (in and out of <figref idref="f0016">Figure 16</figref>), but the invention is not so limited. The illuminator 196 is also attached to the camera platform 207 such that the illuminator 196 maintains a fixed geometric relationship to the video camera 198. Thus, when the video camera 198 is adjusted to the patient's eye 192, the illuminator 196 is automatically adjusted to the patient's eye 192 in the same regard as well. This may be important to enforce a desired distance (d) and angle of illumination (Φ) of the patient's eye 192, as illustrated in <figref idref="f0016">Figure 16</figref>, to properly capture the interference interactions of the specularly reflected light from the patient's tear film at the proper angle of incidence according to Snell's law, since the OSI device 170 is programmed to assume a certain distance and certain angles of incidence. In the OSI device 170 in <figref idref="f0016">Figure 16</figref>, the angle of illumination (Φ) of the patient's eye 192 relative to the camera 198 axis is approximately 30 degrees at the center of the illuminator 196 and includes a relatively large range of angles from about 5 to 60 degrees, but any angle may be provided.

0066<figref idref="f0017 f0018 f0019 f0020">Figures 17-20</figref> provide more detail on the illuminator 196. As illustrated in <figref idref="f0017">Figure 17</figref>, the exemplary illuminator 196 is provided on an arced surface 208 (see also, <figref idref="f0017 f0018">Figures 17-18</figref>) of approximately 75 degrees to provide a large area, broad spectrum light source covering the visible regions of approximately 400 nanometers (nm) to 700 nm. In this embodiment, the arced surface 208 has a radius to an imaginary center of approximately 190 mm ("r" in <figref idref="f0017">Figure 17</figref>) and has a face 250 mm high by 100 mm wide. The arced surface 208 could be provided as a flat surface, but an arced surface may allow for: better illumination uniformity, uniform tile sizes, a smaller sized illuminator 196 for packaging constraints, while providing the same effective illumination area capability. In this example, the illuminator 196 is a Lambertian emitter wherein the light emitter has approximately the same intensity in all directions; however, the present invention is not so limited. The illuminator 196 is arranged so that, from the perspective of the camera 198, emitted light rays are specularly reflected from the tear film of the patient's eye 192 and undergo constructive and destructive interference in the lipid layer and layers beneath the lipid layer. In this embodiment, the illuminator 196 is comprised of high efficiency, white light emitting diodes (LEDs) 210 (see <figref idref="f0017">Figures 17</figref> and <figref idref="f0018">18</figref>) mounted on a printed circuit board (PCB) 212 (<figref idref="f0018">Figure 18</figref>), wherein each LED 210 or each grouping of LEDs is independently addressable by the control system to be turned on and off, which will be used when providing a tiled illumination approach of the patient's tear film. Supporting circuitry (not shown) may be included to control operation of the LEDs 210, and to automatically shut off the LEDs 210 when the OSI device 170 is not in use. Each LED 210 has a 120 degree ("Lambertian") forward projection angle, a 1350 mcd maximum intensity, manufactured by LEDtronics. Other light sources other than LEDs are also possible, including but not limited to lasers, incandescent light, and organic LEDs (OLEDs), as examples. Further, the light source is not required to be a Lambertian emitter. For example, the light emitted from the light source may be collimated.

0067As illustrated in <figref idref="f0019">Figure 19</figref>, the PCB 212 is placed inside an illuminator housing 214. The illuminator housing 214 is comprised of two side panels 216A, 216B that are disposed on opposite sides of the arced surfaced 208 when held by base and top panels 218, 220, and also includes a rear panel 222. The arced surface 208 is comprised of a diffuser 209 to diffuse the light emitted by the LEDs 210. The diffuser 208 can be selected to minimize intensity reduction, while providing sufficient scattering to make the illumination uniform light wave fall off on the light emitted by the outside LEDs 210. The diffuser 209, PCB 212, and rear panel 222 are flexible and fit within grooves 223 located in the top and base panels 220, 218, and grooves 224 located in the side panels 216A, 216B. The illuminator housing 214 is snapped together and the side panels 216A, 216B are then screwed to the top and base panels 220, 218.

0068The diffuser 209 may also be comprised of more than one diffuser panel to improve uniformity in the light emitted from the illuminator 196. The side panels 216A, 216B and the base and top panels 218, 220 form baffles around the PCB 212 and the LEDs 210. The inside of these surfaces may contain a reflective film (e.g., 3M ESR film) to assist in the uniformity of light emitted by the LEDs 210. The reflective film may assist in providing a uniform light intensity over an entire area or region of interest on a patient's tear film. This may be particularly an issue on the outer edges of the illumination pattern. If a tiled approach is employed to illuminate a patient's tear film, whereby only a subset of the LEDs 210 within baffle partitions in the illuminator 196 are turned on at one time, additional edges will be formed as opposed to a single outer edge if all LEDs 210 are turned on with no tile baffles. The baffle partitions are used to delineate individual tiles and form sharp illumination interaction definition between tiles. The fall off of light intensity at the outer edges of the illumination interaction or at tile partition edges may be controlled to be between approximately three percent (3%) and seven percent (7%). The diffuser 209 should also be sufficiently tightly held to the edges and to the tile baffles in the illuminator housing 214 to prevent or reduce shadows on in the illumination pattern.

0069Providing individually controllable LEDs 210 in the illuminator 196 facilitates providing the tiled pattern illumination previously described. In this manner, certain groupings of LEDs 210 can be controlled to be turned on and off to provide a desired tiled illumination of the patient's tear film. <figref idref="f0020 f0021">Figures 20-24</figref> show several exemplary arrangements of organizing the control of the LEDs 210 into groupings to provide tiled illumination of a tear film by the illuminator 196 in the OSI device 170. In <figref idref="f0020">Figure 20</figref>, the LEDs 210 in the illuminator 196 are divided up into two groups (labeled 1-2) of tiles 230 each having a 4x6 array of LEDs 210. In this manner, the PCB 212 contains two hundred eighty-eight (288) LEDs 210. The groups are provided ideally to provide uniform diffuse illumination from the illuminator 196 to capture background signal in the form of diffuse illumination from the illuminator 196 in images of the patient's tear film, as previously described. First, the LEDs 210 in the tiles 230 provided in group 1 are illuminated in a first mode and a first image of the patient's tear film is captured. Then, group 2 is illuminated in a second mode and a second image is captured. This process can be repeated alternating lighting modes between groups 1 and 2 to obtain a time-based sequence of images. The first and second images can then be combined to eliminate or reduce background signal in the interference signal representing the specularly reflected light from the tear film, as previously discussed. For example, in order to maintain an overall frame rate of thirty (30) fps, the video camera 198 would have to operate in at least 60 fps (30 fps x 2 groupings).

0070Other groups are also possible. <figref idref="f0020">Figure 21</figref> provides four groupings (labeled 1-4), with each group perhaps having a 4x6 array of LEDs 210. The LEDs 210 in each group are illuminated one at a time in sequence (i.e., group 1, 2, 3, 4, 1, etc.) and an image is taken of the patient's tear film, with all images composed together to provide an illuminated, background signal reduced or eliminated, image of the patient's tear film. <figref idref="f0020">Figure 22</figref> also provides four groupings (labeled 1-4), with each group having an array of LEDs 210. In order to maintain an overall frame rate of fifteen (15) fps, the video camera 198 would have to operate in at least 60 fps (15 fps x 4 groupings). The groupings arranged so each group provides, as similar as possible, the same average illumination geometry to the subject's eye.

0071<figref idref="f0021">Figure 23</figref> provides twelve groupings (labeled 1-12), with each group also having an array of LEDs 210. In order to maintain an overall frame rate of fifteen (15) fps, the video camera 198 would have to operate at 180 fps (15 fps x 12 groupings). A high-speed complementary metal oxide (CMOS) camera may be employed as opposed to a CCD camera to achieve this frame rate. <figref idref="f0021">Figure 24</figref> also provides twelve groupings (labeled 1-12), with each group having a 3x4 array of LEDs 210. (Higher number of groups provides the advantage of lowering the background image level due to the illuminator relative to the specular image, thus improving the ability to remove the induced background. Working against the advantage, higher numbers of tile groups can make it more difficult to produce the same average illumination geometry for all tile modes. Fortunately, with enough tile groups, we may be able to ignore the background contribution from the illuminator light entirely, but the ambient and stray light may need subtraction by some means. In the limit, increasing the number of groups begins to approach a point to point scanning system.)

<u>System Level</u>

0072Now that the imaging and illumination functions of the OSI device 170 have been described, <figref idref="f0022">Figure 25A</figref> illustrates a system level diagram illustrating more detail regarding the control system and other internal components of the OSI device 170 provided inside the housing 172 according to one embodiment to capture images of a patient's tear film and process those images. As illustrated therein, a control system 240 is provided that provides the overall control of the OSI device 170. The control system 240 may be provided by any microprocessor-based or computer system. The control system 240 illustrated in <figref idref="f0022">Figure 25A</figref> is provided in a system-level diagram and does not necessarily imply a specific hardware organization and/or structure. As illustrated therein, the control system 240 contains several systems. A camera settings system 242 may be provided that accepts camera settings from a clinician user. Exemplary camera settings 244 are illustrated, but may be any type according to the type and model of camera provided in the OSI device 170 as is well understood by one of ordinary skill in the art.

0073The camera settings 244 may be provided to (The Imaging Source) camera drivers 246, which may then be loaded into the video camera 198 upon initialization of the OSI device 170 for controlling the settings of the video camera 198. The settings and drivers may be provided to a buffer 248 located inside the video camera 198 to store the settings for controlling a CCD 250 for capturing ocular image information from a lens 252. Ocular images captured by the lens 252 and the CCD 250 are provided to a de-Bayering function 254 which contains an algorithm for post-processing of raw data from the CCD 250 as is well known. The ocular images are then provided to a video acquisition system 256 in the control system 240 and stored in memory, such as random access memory (RAM) 258. The stored ocular images or signal representations can then be provided to a pre-processing system 260 and a post-processing system 262 to manipulate the ocular images to obtain the interference interactions of the specularly reflected light from the tear film and analyze the information to determine characteristics of the tear film. Pre-processing settings 264 and post-processing settings 266 can be provided to the pre-processing system 260 and post-processing system 262, respectively, to control these functions. These settings 264, 266 will be described in more detail below. The post-processed ocular images and information may also be stored in mass storage, such as disk memory 268, for later retrieval and viewing on the display 174.

0074The control system 240 may also contain a visualization system 270 that provides the ocular images to the display 174 to be displayed in human-perceptible form on the display 174. Before being displayed, the ocular images may have to be pre-processed in a pre-processing video function 272. For example, if the ocular images are provided by a linear camera, non-linearity (i.e. gamma correction) may have to be added in order for the ocular images to be properly displayed on the display 174. Further, contrast and saturation display settings 274, which may be controlled via the display 174 or a device communicating to the display 174, may be provided by a clinician user to control the visualization of ocular images displayed on the display 174. The display 174 is also adapted to display analysis result information 276 regarding the patient's tear film, as will be described in more detail below. The control system 240 may also contain a user interface system 278 that drives a graphical user interface (GUI) utility 280 on the display 174 to receive user input 282. The user input 282 can include any of the settings for the OSI device 170, including the camera settings 244, the pre-processing settings 264, the post-processing settings 266, the display settings 274, the visualization system 270 enablement, and video acquisition system 256 enablement, labeled 1-6. The GUI utility 280 may only be accessible by authorized personnel and used for calibration or settings that would normally not be changed during normal operation of the OSI device 170 once configured and calibrated.

<u>Overall Process Flow</u>

0075<figref idref="f0023">Figure 25B</figref> illustrates an exemplary overall flow process performed by the OSI device 170 for capturing tear film images from a patent and analysis for TFLT measurement. As illustrated in <figref idref="f0023">Figure 25B</figref>, the video camera 198 is connected via a USB port 283 to the control system 240 (see <figref idref="f0022">Figure 25A</figref>) for control of the video camera 198 and for transferring images of a patient's tear film taken by the video camera 198 back to the control system 240. The control system 240 includes a compatible camera driver 246 to provide a transfer interface between the control system 240 and the video camera 198. Prior to tear film image capture, the configuration or camera settings 244 are loaded into the video camera 198 over the USB port 283 to prepare the video camera 198 for tear film image capture (block 285). Further, an audio video interleaved (AVI) container is created by the control system 240 to store video of tear film images to be captured by the video camera 198 (block 286). At this point, the video camera 198 and control system 240 are ready to capture images of a patient's tear film. The control system 240 waits for a user command to initiate capture of a patient's tear film (blocks 287, 288).

0076Once image capture is initiated (block 288), the control system enables image capture to the AVI container previously setup (block 286) for storage of images captured by the video camera 198 (block 289). The control system 240 controls the video camera 198 to capture images of the patient's tear film (block 289) until timeout or the user terminates image capture (block 290) and image capture halts or ends (block 291). Images captured by the video camera 198 and provided to the control system 240 over the USB port 283 are stored by the control system 240 in RAM 268.

0077The captured images of the patient's ocular tear film can subsequently be processed and analyzed to perform TFLT measurement, as described in more detail below and throughout the remainder of this disclosure. The process in this embodiment involves processing tear film image pairs to perform background subtraction, as previously discussed. For example, image tiling may be performed to provide the tear film image pairs, if desired. The processing can include simply displaying the patient's tear film or performing TFLT measurement (block 293). If the display option is selected to allow a technician to visually view the patient's tear film, display processing is performed (block 294) which can be the display processing 270 described in more detail below with regard to <figref idref="f0032">Figure 34</figref>. For example, the control system 240 can provide a combination of images of the patient's tear film that show the entire region of interest of the tear film on the display 174. The displayed image may include the background signal or may have the background signal subtracted. If TFLT measurement is desired, the control system 240 performs pre-processing of the tear film images for TFLT measurement (block 295), which can be the pre-processing 260 described in more detail below with regard to <figref idref="f0024">Figure 26</figref>. The control system 240 also performs post-processing of the tear film images for TFLT measurement (block 296), which can be the post-processing 262 described in more detail below with regard to <figref idref="f0036">Figure 36</figref>.

<u>Pre-Processing</u>

0078<figref idref="f0024">Figure 26</figref> illustrates an exemplary pre-processing system 260 for pre-processing ocular tear film images captured by the OSI device 170 for eventual analysis and TFLT measurement. In this system, the video camera 198 has already taken the first and second tiled images of a patient's ocular tear film, as previously illustrated in <figref idref="f0011">Figures 11A and 11B</figref>, and provided the images to the video acquisition system 256. The frames of the first and second images were then loaded into RAM 258 by the video acquisition system 256. Thereafter, as illustrated in <figref idref="f0024">Figure 26</figref>, the control system 240 commands the pre-processing system 260 to pre-process the first and second images. An exemplary GUI utility 280 is illustrated in <figref idref="f0025">Figure 27</figref> that may be employed by the control system 240 to allow a clinician to operate the OSI device 170 and control pre-processing settings 264 and post-processing settings 266, which will be described later in this application. In this regard, the pre-processing system 260 loads the first and second image frames of the ocular tear film from RAM 258 (block 300). The exemplary GUI utility 280 in <figref idref="f0025">Figure 27</figref> allows for a stored image file of previously stored video sequence of first and second image frames captured by the video camera 198 by entering a file name in the file name field 351. A browse button 352 also allows searches of the memory for different video files, which can either be buffered by selecting a buffered box 354 or loaded for pre-processing by selecting the load button 356.

0079If the loaded first and second image frames of the tear film are buffered, they can be played using display selection buttons 358, which will in turn display the images on the display 174. The images can be played on the display 174 in a looping fashion, if desired, by selecting the loop video selection box 360. A show subtracted video selection box 370 in the GUI utility 280 allows a clinician to show the resulting, subtracted video images of the tear film on the display 174 representative of the resulting signal comprised of the second output signal combined or subtracted from the first output signal, or vice versa. Also, by loading the first and second image frames, the previously described subtraction technique can be used to remove background image from the interference signal representing interference of the specularly reflected light from the tear film, as previously described above and illustrated in <figref idref="f0012">Figure 12</figref> as an example. The first image is subtracted from the second image to subtract or remove the background signal in the portions producing specularly reflected light in the second image, and vice versa, and then combined to produce an interference interaction of the specularly reflected light of the entire area or region of interest of the tear film, as previously illustrated in <figref idref="f0012">Figure 12</figref> (block 302 in <figref idref="f0024">Figure 26</figref>). For example, this processing could be performed using the Matlab® function "cvAbsDiff."

0080The subtracted image containing the specularly reflected light from the tear film can also be overlaid on top of the original image capture of the tear film to display an image of the entire eye and the subtracted image in the display 174 by selecting the show overlaid original video selection box 362 in the GUI utility 280 of <figref idref="f0025">Figure 27</figref>. An example of an overlaid original video to the subtracted image of specularly reflected light from the tear film is illustrated in the image 363 of <figref idref="f0026">Figure 28</figref>. This overlay is provided so that flashing images of specularly reflected light from the tear film are not displayed, which may be unpleasant to visualize. The image 363 of the tear film illustrated in <figref idref="f0026">Figure 28</figref> was obtained with a DBK 21AU04 Bayer VGA (640x480) video camera having a Pentax VS-LD25 Daitron 25-mm fixed focal length lens with maximum aperture at a working distance of 120 mm and having the following settings, as an example: <ul id="ul0002" list-style="none" compact="compact"><li>Gamma = 100 (to provide linearity with exposure value)</li><li>Exposure = 1/16 second</li><li>Frame rate = 60 fps</li><li>Data Format = BY8</li><li>Video Format =- uncompressed, RGB 24-bit AVI</li><li>Hue = 180 (neutral, no manipulation)</li><li>Saturation = 128(neutral, no manipulation)</li><li>Brightness = 0 (neutral, no manipulation)</li><li>Gain = 260 (minimum available setting in this camera driver)</li><li>White balance = B=78; R=20.</li></ul>

Thresholding

0081Any number of optional pre-processing steps and functions can next be performed on the resulting combined tear film image(s), which will now be described. For example, an optional threshold pre-processing function may be applied to the resulting image or each image in a video of images of the tear film (e.g., <figref idref="f0012">Figure 12</figref>) to eliminate pixels that have a subtraction difference signal below a threshold level (block 304 in <figref idref="f0024">Figure 26</figref>). Image threshold provides a black and white mask (on/off) that is applied to the tear film image being processed to assist in removing residual information that may not be significant enough to be analyzed and/or may contribute to inaccuracies in analysis of the tear film. The threshold value used may be provided as part of a threshold value setting provided by a clinician as part of the pre-processing settings 264, as illustrated in the system diagram of <figref idref="f0022">Figure 25A</figref>. For example, the GUI utility 280 in <figref idref="f0025">Figure 27</figref> includes a compute threshold selection box 372 that may be selected to perform thresholding, where the threshold brightness level can be selected via the threshold value slide 374. The combined tear film image of <figref idref="f0012">Figure 12</figref> is copied and converted to grayscale. The grayscale image has a threshold applied according to the threshold setting to obtain a binary (black/white) image that will be used to mask the combined tear film image of <figref idref="f0012">Figure 12</figref>. After the mask is applied to the combined tear film image of <figref idref="f0012">Figure 12</figref>, the new combined tear film image is stored in RAM 258. The areas of the tear film image that do not meet the threshold brightness level are converted to black as a result of the threshold mask.

0082<figref idref="f0027">Figures 29A and 29B</figref> illustrate examples of threshold masks for the combined tear film provided in <figref idref="f0012">Figure 12</figref>. <figref idref="f0027">Figure 29A</figref> illustrates a threshold mask 320 for a threshold setting of 70 counts out of a full scale level of 255 counts. <figref idref="f0027">Figure 29B</figref> illustrates a threshold mask 322 for a threshold setting of 50. Note that the threshold mask 320 in <figref idref="f0027">Figure 29A</figref> contains less portions of the combined tear film image, because the threshold setting is higher than for the threshold mask 322 of <figref idref="f0027">Figure 29B</figref>. When the threshold mask according to a threshold setting of 70 is applied to the exemplary combined tear film image of <figref idref="f0012">Figure 12</figref>, the resulting tear film image is illustrated <figref idref="f0028">Figure 30</figref>. Much of the residual subtracted background image that surrounds the area or region of interest has been masked away.

Erode and Dilate

0083Another optional pre-processing function that may be applied to the resulting image or each image in a video of images of the tear film to correct anomalies in the combined tear film image(s) is the erode and dilate functions (block 306 in <figref idref="f0024">Figure 26</figref>). The erode function generally removes small anomaly artifacts by subtracting objects with a radius smaller than an erode setting (which is typically in number of pixels) removing perimeter pixels where interference information may not be as distinct or accurate. The erode function may be selected by a clinician in the GUI utility 280 (see <figref idref="f0025">Figure 27</figref>) by selecting the erode selection box 376. If selected, the number of pixels for erode can be provided in an erode pixels text box 378. Dilating generally connects areas that are separated by spaces smaller than a minimum dilate size setting by adding pixels of the eroded pixel data values to the perimeter of each image object remaining after the erode function is applied. The dilate function may be selected by a clinician in the GUI utility 280 (see <figref idref="f0025">Figure 27</figref>) by providing the number of pixels for dilating in a dilate pixels text box 380. Erode and dilate can be used to remove small region anomalies in the resulting tear film image prior to analyzing the interference interactions to reduce or avoid inaccuracies. The inaccuracies may include those caused by bad pixels of the video camera 198 or from dust that may get onto a scanned image, or more commonly, spurious specular reflections such as: tear film meniscus at the juncture of the eyelids, glossy eyelash glints, wet skin tissue, etc. <figref idref="f0029">Figure 31</figref> illustrates the resulting tear film image of <figref idref="f0028">Figure 30</figref> after erode and dilate functions have been applied and the resulting tear film image is stored in RAM 258. As illustrated therein, pixels previously included in the tear film image that were not in the tear film area or region of interest are removed. This prevents data in the image outside the area or region of interest from affecting the analysis of the resulting tear film image(s).

Removing Blinks / Other Anomalies

0084Another optional pre-processing function that may be applied to the resulting image or each image in a video of images of the tear film to correct anomalies in the resulting tear film image is to remove frames from the resulting tear film image that include patient blinks or significant eye movements (block 308 in <figref idref="f0024">Figure 26</figref>). As illustrated in <figref idref="f0024">Figure 26</figref>, blink detection is shown as being performed after a threshold and erode and dilate functions are performed on the tear film image or video of images. Alternatively, the blink detection could be performed immediately after background subtraction, such that if a blink is detected in a given frame or frames, the image in such frame or frames can be discarded and not pre-processed. Not pre-processing images where blinks are detected may increase the overall speed of pre-processing. The remove blinks or movement pre-processing may be selectable. For example, the GUI utility 280 in <figref idref="f0025">Figure 27</figref> includes a remove blinks selection box 384 to allow a user to control whether blinks and/or eye movements are removed from a resulting image or frames of the patient's tear film prior to analysis. Blinking of the eyelids covers the ocular tear film, and thus does not produce interference signals representing specularly reflected light from the tear film. If frames containing whole or partial blinks obscuring the area or region of interest in the patient's tear film are not removed, it would introduce errors in the analysis of the interference signals to determine characteristics of the TFLT of the patient's ocular tear film. Further, frames or data with significant eye movement between sequential images or frames can be removed during the detect blink pre-processing function. Large eye movements could cause inaccuracy in analysis of a patient's tear film when employing subtraction techniques to remove background signal, because subtraction involves subtracting frame-pairs in an image that closely match spatially. Thus, if there is significant eye movement between first and second images that are to be subtracted, frame pairs may not be closely matched spatially thus inaccurately removing background signal, and possibly removing a portion of the interference image of specularly reflected light from the tear film.

0085Different techniques can be used to determine blinks in an ocular tear film image and remove the frames as a result. For example, in one embodiment, the control system 240 directs the pre-processing system 260 to review the stored frames of the resulting images of the tear film to monitor for the presence of an eye pupil using pattern recognition. A Hough Circle Transform may be used to detect the presence of the eye pupil in a given image or frame. If the eye pupil is not detected, it is assembled such that the image or frame contains an eye blink and thus should be removed or ignored during pre-processing from the resulting image or video of images of the tear film. The resulting image or video of images can be stored in RAM 258 for subsequent processing and/or analyzation.

0086In another embodiment, blinks and significant eye movements are detected using a histogram sum of the intensity of pixels in a resulting subtracted image or frame of a first and second image of the tear film. An example of such a histogram 329 is illustrated in <figref idref="f0030">Figure 32</figref>. The resulting or subtracted image can be converted to grayscale (i.e., 255 levels) and a histogram generated with the gray levels of the pixels. In the histogram 329 of <figref idref="f0030">Figure 32</figref>, the x-axis contains gay level ranges, and the number of pixels falling within each gray level is contained in the y-axis. The total of all the histogram 329 bins are summed. In the case of two identical frames that are subtracted, the histogram sum would be zero. However, even without an eye blink or significant eye movement, two sequentially captured frames of the patient's eye and the interference signals representing the specularly reflected light from the tear film are not identical. However, frame pairs with little movement will have a low histogram sum, while frame pairs with greater movement will yield a larger histogram sum. If the histogram sum is beyond a pre-determined threshold, an eye blink or large eye movement can be assumed and the image or frame removed. For example, the GUI utility 280 illustrated in <figref idref="f0025">Figure 27</figref> includes a histogram sum slide bar 386 that allows a user to set the threshold histogram sum. The threshold histogram sum for determining whether a blink or large eye movement should be assumed and thus the image removes from analysis of the patient's tear film can be determined experimentally, or adaptively over the course of a frame playback, assuming that blinks occur at regular intervals.

0087An advantage of a histogram sum of intensity method to detect eye blinks or significant eye movements is that the calculations are highly optimized as opposed to pixel-by-pixel analysis, thus assisting with real-time processing capability. Further, there is no need to understand the image structure of the patient's eye, such as the pupil or the iris details. Further, the method can detect both blinks and eye movements.

0088Another alternate technique to detect blinks in the tear film image or video of images for possible removal is to calculate a simple average gray level in an image or video of images. Because the subtracted, resulting images of the tear film subtract background signal, and have been processed using a threshold mask, and erode and dilate functions performed in this example, the resulting images will have a lower average gray level due to black areas present than if a blink is present. A blink contains skin color, which will increase the average gray level of an image containing a blink. A threshold average gray level setting can be provided. If the average gray level of a particular frame is below the threshold, the frame is ignored from further analysis or removed from the resulting video of frames of the tear film.

0089Another alternate technique to detect blinks in an image or video of images for removal is to calculate the average number of pixels in a given frame that have a gray level value below a threshold gray level value. If the percentage of pixels in a given frame is below a defined threshold percentage, this can be an indication that a blink has occurred in the frame, or that the frame is otherwise unworthy of consideration when analyzing the tear film. Alternatively, a spatial frequency calculation can be performed on a frame to determine the amount of fine detail in a given frame. If the detail present is below a threshold detail level, this may be an indication of a blink or other obscurity of the tear film, since skin from the eyelid coming down and being captured in a frame will have less detail than the subtracted image of the tear film. A histogram can be used to record any of the above-referenced calculations to use in analyzing whether a given frame should be removed from the final pre-processed resulting image or images of the tear film for analyzation.

ICC Profiling

0090Pre-processing of the resulting tear film image(s) may also optionally include applying an International Colour Consortium (ICC) profile to the pre-processed interference images of the tear film (block 310, <figref idref="f0024">Figure 26</figref>). <figref idref="f0031">Figure 33</figref> illustrates an optional process of loading an ICC profile into an ICC profile 331 in the control system 240 (block 330). In this regard, the GUI utility 280 illustrated in <figref idref="f0025">Figure 27</figref> also includes an apply ICC box 392 that can be selected by a clinician to load the ICC profile 331. The ICC profile 331 may be stored in memory in the control system 240, including in RAM 258. In this manner, the GUI utility 280 in <figref idref="f0025">Figure 27</figref> also allows for a particular ICC profile 331 to be selected for application in the ICC profile file text box 394. The ICC profile 331 can be used to adjust color reproduction from scanned images from cameras or other devices into a standard red-green-blue (RGB) color space (among other selectable standard color spaces) defined by the ICC and based on a measurement system defined internationally by the Commission Internationale de l'Eclairage (CIE). Adjusting the pre-processed resulting tear film interference images corrects for variations in the camera color response and the light source spectrum and allows the images to be compatibly compared with a tear film layer interference model to measure the thickness of a TFLT, as will be described later in this application. The tear film layers represented in the tear film layer interference model can be LLTs, ALTs, or both, as will be described in more detail below.

0091In this regard, the ICC profile 331 may have been previously loaded to the OSI device 170 before imaging of a patient's tear film and also applied to a tear film layer interference model when loaded into the OSI device 170 independent of imaging operations and flow. As will be discussed in more detail below, a tear film layer interference model in the form of a TFLT palette 333 containing color values representing interference interactions from specularly reflected light from a tear film for various LLTs and ALTs can also be loaded into the OSI device 170 (block 332 in <figref idref="f0036">Figure 36</figref>). The tear film layer interference model 333 contains a series of color values that are assigned LLTs and/or ALTs based on a theoretical tear film layer interference model to be compared against the color value representations of interference interactions in the resulting image(s) of the patient's tear film. When applying the optional ICC profile 331 to the tear film layer interference model 333 (block 334 in <figref idref="f0031">Figure 33</figref>), the color values in both the tear film layer interference model and the color values representing interference interactions in the resulting image of the tear film are adjusted for a more accurate comparison between the two to measure LLT and/or ALT.

Brightness

0092Also as an optional pre-processing step, brightness and red-green-blue (RGB) subtract functions may be applied to the resulting interference signals of the patient's tear film before post-processing for analysis and measuring TFLT is performed (blocks 312 and 314 in <figref idref="f0024">Figure 26</figref>). The brightness may be adjusted pixel-by-pixel by selecting the adjust brightness selection box 404 according to a corresponding brightness level value provided in a brightness value box 406, as illustrated in the GUI utility 280 of <figref idref="f0025">Figure 27</figref>. When the brightness value box 406 is selected, the brightness of each palette value of the tear film interference model 333 is also adjusted accordingly.

RGB Subtraction (Normalization)

0093The RGB subtract function subtracts a DC offset from the interference signal in the resulting image(s) of the tear film representing the interference interactions in the interference signal. An RGB subtract setting may be provided from the pre-processing settings 264 to apply to the interference signal in the resulting image of the tear film to normalize against. As an example, the GUI utility 280 in <figref idref="f0025">Figure 27</figref> allows an RGB offset to be supplied by a clinician or other technician for use in the RGB subtract function. As illustrated therein, the subtract RGB function can be activated by selecting the RGB subtract selection box 396. If selected, the individual RGB offsets can be provided in offset value input boxes 398. After pre-processing is performed, if any, on the resulting image, the resulting image can be provided to a post-processing system to measure TLFT (block 316), as discussed later below in this application.

Displaying Images

0094The resulting images of the tear film may also be displayed on the display 174 of the OSI device 170 for human diagnosis of the patient's ocular tear film. The OSI device 170 is configured so that a clinician can display and see the raw captured image of the patient's eye 192 by the video camera 198, the resulting images of the tear film before pre-processing, or the resulting images of the tear film after pre-processing. Displaying images of the tear film on the display 174 may entail different settings and steps. For example, if the video camera 198 provides linear images of the patient's tear film, the linear images must be converted into a non-linear format to be properly displayed on the display 174. In this regard, a process that is performed by the visualization system 270 according to one embodiment is illustrated in <figref idref="f0032">Figure 34</figref>.

0095As illustrated in <figref idref="f0032">Figure 34</figref>, the video camera 198 has already taken the first and second tiled images of a patient's ocular tear film as previously illustrated in <figref idref="f0011">Figures 11A and 11B</figref>, and provided the images to the video acquisition system 256. The frames of the first and second images were then loaded into RAM 258 by the video acquisition system 256. Thereafter, as illustrated in <figref idref="f0032">Figure 34</figref>, the control system 240 commands the visualization system 270 to process the first and second images to prepare them for being displayed on the display 174, 338. In this regard, the visualization system 270 loads the first and second image frames of the ocular tear film from RAM 258 (block 335). The previously described subtraction technique is used to remove background signal from the interference interactions of the specularly reflected light from the tear film, as previously described above and illustrated in <figref idref="f0012">Figure 12</figref>. The first image(s) is subtracted from the second image(s) to remove background signal in the illuminated portions of the first image(s), and vice versa, and the subtracted images are then combined to produce an interference interaction of the specularly reflected light of the entire area or region of interest of the tear film, as previously discussed and illustrated in <figref idref="f0012">Figure 12</figref> (block 336 in <figref idref="f0032">Figure 34</figref>).

0096Again, for example, this processing could be performed using the Matlab® function "cvAbsDiff." Before being displayed, the contrast and saturation levels for the resulting images can be adjusted according to contrast and saturation settings provided by a clinician via the user interface system 278 and/or programmed into the visualization system 270 (block 337). For example, the GUI utility 280 in <figref idref="f0025">Figure 27</figref> provides an apply contrast button 364 and a contrast setting slide 366 to allow the clinician to set the contrast setting in the display settings 274 for display of images on the display 174. The GUI utility 280 also provides an apply saturation button 368 and a saturation setting slide 369 to allow a clinician to set the saturation setting in the display settings 274 for the display of images on the display 174. The images can then be provided by the visualization system 270 to the display 174 for displaying (block 338 in <figref idref="f0032">Figure 34</figref>). Also, any of the resulting images after pre-processing steps in the pre-processing system 260 can be provided to the display 174 for processing.

0097<figref idref="f0033 f0034 f0035">Figures 35A-35C</figref> illustrate examples of different tear film images that are displayed on the display 174 of the OSI device 170. <figref idref="f0033">Figure 35A</figref> illustrates a first image 339 of the patient's tear film showing the tiled pattern captured by the video camera 198. This image is the same image as illustrated in <figref idref="f0011">Figure 11A</figref> and previously described above, but processed from a linear output from the video camera 198 to be properly displayed on the display 174. <figref idref="f0034">Figure 35B</figref> illustrates a second image 340 of the patient's tear film illustrated in <figref idref="f0011">Figure 11B</figref> and previously described above. <figref idref="f0035">Figure 35C</figref> illustrates a resulting "overlaid" image 341 of the first and second images 339, 340 of the patient's tear film and to provide interference interactions of the specularly reflected light from the tear film over the entire area or region of interest. This is the same image as illustrated in <figref idref="f0012">Figure 12</figref> and previously described above.

0098In this example, the original number of frames of the patient's tear film captured can be reduced by half due to the combination of the first and second tiled pattern image(s). Further, if frames in the subtracted image frames capture blinks or erratic movements, and these frames are eliminated in pre-processing, a further reduction in frames will occur during pre-processing from the number of images raw captured in images of the patient's tear film. Although these frames are eliminated from being further processed, they can be retained for visualization rendering a realistic and natural video playback..Further, by applying a thresholding function and erode and dilating functions, the number of non-black pixels which contain TLFT interference information is substantially reduced as well. Thus, the amount of pixel information that is processed by the post-processing system 262 is reduced, and may be on the order of 70% less information to process than the raw image capture information, thereby pre-filtering for the desired interference ROI and reducing or elimination potentially erroneous information as well as allowing for faster analysis due to the reduction in information.

0099At this point, the resulting images of the tear film have been pre-processed by the pre-processing system 260 according to whatever pre-processing settings 264 and pre-processing steps have been selected or implemented by the control system 240. The resulting images of the tear film are ready to be processed for analyzing and determining TFLT. In this example, this is performed by the post-processing system 262 in <figref idref="f0022">Figure 25A</figref> and is based on the post-processing settings 266 also illustrated therein. An embodiment of the post-processing performed by the post-processing system 262 is illustrated in the flowchart of <figref idref="f0036">Figure 36</figref>.

<u>Tear Film Interference Models</u>

0100As illustrated in <figref idref="f0036">Figure 36</figref>, pre-processed images 343 of the resulting images of the tear film are retrieved from RAM 258 where they were previously stored by the pre-processing system 260. Before discussing the particular embodiment of the post-processing system 262 in <figref idref="f0036">Figure 36</figref>, in general, to measure TFLT, the RGB color values of the pixels in the resulting images of the tear film are compared against color values stored in a tear film interference model that has been previously loaded into the OSI device 170 (see <figref idref="f0031">Figure 33</figref>. The tear film interference model may be stored as a TFLT palette 333 containing RGB values representing interference colors for given LLTs and/or ALTs. The TFLT palette contains interference color values that represent TFLTs based on a theoretical tear film interference model in this embodiment. Depending on the TFLT palette provided, the interference color values represented therein may represent LLTs, ALTs, or both. An estimation of TFLT for each ROI pixel is based on this comparison. This estimate of TFLT is then provided to the clinician via the display 174 and/or recorded in memory to assist in diagnosing DES.

0101Before discussing embodiments of how the TFLTs are estimated from the pre-processed resulting image colored interference interactions resulting from specularly reflected light from the tear film, tear film interference modeling is first discussed. Tear film interference modeling can be used to determine an interference color value for a given TFLT to measure TFLT, which can include both LLT and/or ALT.

0102Although the interference signals representing specularly reflected light from the tear film are influenced by all layers in the tear film, the analysis of interference interactions due to the specularly reflected light can be analyzed under a 2-wave tear film model (i.e., two reflections) to measure LLT. A 2-wave tear film model is based on a first light wave(s) specularly reflecting from the air-to-lipid layer transition of a tear film and a second light wave specularly reflecting from the lipid layer-to-aqueous layer transition of the tear film. In the 2-wave model, the aqueous layer is effective ignored and treated to be of infinite thickness. To measure LLT using a 2-wave model, a 2-wave tear film model was developed wherein the light source and lipid layers of varying thicknesses were modeled mathematically. To model the tear-film interference portion, commercially available software, such as that available by FilmStar and Zemax as examples, allows image simulation of thin films for modeling. Relevant effects that can be considered in the simulation include refraction, reflection, phase difference, polarization, angle of incidence, and refractive index wavelength dispersion. For example, a lipid layer could be modeled as having an index of refraction of 1.48 or as a fused silica substrate (SiO₂) having a 1.46 index of refraction. A back material, such as Magnesium Flouride (MgF₂) having an index of refraction of 1.38 may be used to provide a 2-wave model of air/SiO₂/MgF₂ (1.0/1.46/1.38). To obtain the most accurate modeling results, the model can include the refractive index and wavelength dispersion values of biological lipid material and biological aqueous material, found from the literature, thus to provide a precise two-wave model of air/lipid/aqueous layers. Thus, a 2-wave tear film interference model allows measurement of LLT regardless of ALT.

0103Simulations can be mathematically performed by varying the LLT between 10 to 300 nm. As a second step, the RGB color values of the resulting interference signals from the modeled light source causing the modeled lipid layer to specularly reflected light and received by the modeled camera were determined for each of the modeled LLT. These RGB color values representing interference interactions in specularly reflected light from the modeled tear film were used to form a 2-wave model LLT palette, wherein each RGB color value is assigned a different LLT. The resulting subtracted image of the first and second images from the patient's tear film containing interference signals representing specularly reflected light are compared to the RGB color values in the 2-wave model LLT palette to measure LLT.

0104In another embodiment, a 3-wave tear film interference model may be employed to estimate LLT. A 3-wave tear film interference model does not assume that the aqueous layer is infinite in thickness. In an actual patient's tear film, the aqueous layer is not infinite. The 3-wave tear film interference model is based on both the first and second reflected light waves of the 2-wave model and additionally light wave(s) specularly reflecting from the aqueous-to-mucin layer and/or cornea transitions. Thus, a 3-wave tear film interference model recognizes the contribution of specularly reflected light from the aqueous-to-mucin layer and/or cornea transition that the 2-wave tear film interference model does not. To estimate LLT using a 3-wave tear film interference model, a 3-wave tear film model was previously constructed wherein the light source and a tear film of varying lipid and aqueous layer thicknesses were mathematically modeled. For example, a lipid layer could be mathematically modeled as a material having an index of refraction of 1.48 or as fused silica substrate (SiO₂), which has a 1.46 index of refraction. Different thicknesses of the lipid layer can be simulated. A fixed thickness aqueous layer (e.g., >= 2 µm) could be mathematically modeled as Magnesium Flouride (MgF₂) having an index of refraction of 1.38. A biological cornea could be mathematically modeled as fused silica with no dispersion, thereby resulting in a 3-wave model of air/SiO₂/MgF₂/SiO₂ (i.e., 1.0/1.46/1.38/1.46 with no dispersion). As before, accurate results are obtained if the model can include the refractive index and wavelength dispersion values of biological lipid material, biological aqueous material, and cornea tissue, found from the literature, thus to provide a precise two-wave model of air/lipid/aqueous/cornea layers. The resulting interference interactions of specularly reflected light from the various LLT values and with a fixed ALT value are recorded in the model and, when combined with modeling of the light source and the camera, will be used to compare against interference from specularly reflected light from an actual tear film to measure LLT and/or ALT.

0105In another embodiment of the OSI device 170 and the post-processing system 262 in particular, a 3-wave tear film interference model is employed to estimate both LLT and ALT. In this regard, instead of providing either a 2-wave theoretical tear film interference model that assumes an infinite aqueous layer thickness or a 3-wave model that assumes a fixed or minimum aqueous layer thickness (e.g., ≥ 2µm), a 3-wave theoretical tear film interference model is developed that provides variances in both LLT and ALT in the mathematical model of the tear film. Again, the lipid layer in the tear film model could be modeled mathematically as a material having an index of refraction of 1.48 or as fused silica substrate (SiO₂) having a 1.46 index of refraction. The aqueous layer could be modeled mathematically as Magnesium Flouride (MgF₂) having an index of refraction of 1.38. A biological cornea could be modeled as fused silica with no dispersion, thereby resulting in a 3-wave model of air/SiO₂/MgF₂/SiO₂ (no dispersion). Once again, the most accurate results are obtained if the model can include the refractive index and wavelength dispersion values of biological lipid material, biological aqueous material, and cornea tissue, found from the literature, thus to provide a precise two-wave model of air/lipid/aqueous/cornea layers. Thus, a two-dimensional (2D) TFLT palette 430 (<figref idref="f0037">Figure 37A</figref>) is produced for analysis of interference interactions from specularly reflected light from the tear film. One dimension of the TFLT palette 430 represents a range of RGB color values each representing a given theoretical LLT calculated by mathematically modeling the light source and the camera and calculating the interference interactions from specularly reflected light from the tear film model for each variation in LLT 434 in the tear film interference model. A second dimension of the TFLT palette 430 represents ALT also calculated by mathematically modeling the light source and the camera and calculating the interference interactions from specularly reflected light from the tear film interference model for each variation in ALT 432 at each LLT value 434 in the tear film interference model.

<u>Post-Processins / TFLT Measurement</u>

0106To measure TFLT, a spectral analysis of the resulting interference signal or image is performed during post-processing to calculate a TFLT. In one embodiment, the spectral analysis is performed by performing a look-up in a tear film interference model to compare one or more interference interactions present in the resulting interference signal representing specularly reflected light from the tear film to the RGB color values in the tear film interference model. In this regard, <figref idref="f0037">Figures 37A</figref> and <figref idref="f0038">37B</figref> illustrate two examples of palette models for use in post-processing of the resulting image having interference interactions from specularly reflected light from the tear film using a 3-wave theoretical tear film interference model developed using a 3-wave theoretical tear film model. In general, an RGB numerical value color scheme is employed in this embodiment, wherein the RGB value of a given pixel from a resulting pre-processed tear film image of a patient is compared to RGB values in the 3-wave tear film interference model representing color values for various LLTs and ALTs in a 3-wave modeled theoretical tear film. The closest matching RGB color is used to determine the LLT and/or ALT for each pixel in the resulting signal or image. All pixels for a given resulting frame containing the resulting interference signal are analyzed in the same manner on a pixel-by-pixel basis. A histogram of the LLT and ALT occurrences is then developed for all pixels for all frames and the average LLT and ALT determined from the histogram (block 348 in <figref idref="f0036">Figure 36</figref>).

0107<figref idref="f0037">Figure 37A</figref> illustrates an exemplary TFLT palette 430 in the form of colors representing the included RGB color values representing interference of specularly reflected light from a 3-wave theoretical tear film model used to compared colors from the resulting image of the patient's tear film to estimate LLT and ALT. <figref idref="f0038">Figure 37B</figref> illustrates an alternative example of a TFLT palette 430' in the form of colors representing the included RGB color values representing interference of specularly reflected light from a 3-wave theoretical tear film model used to compare colors from the resulting image of the patient's tear film to estimate LLT and ALT. As illustrated in <figref idref="f0037">Figure 37A</figref>, the TFLT palette 430 contains a plurality of hue colors arranged in a series of rows 432 and columns 434. In this example, there are 144 color hue entries in the palette 430, with nine (9) different ALTs and sixteen (16) different LLTs in the illustrated TFLT palette 430, although another embodiment includes thirty (30) different LLTs. Providing any number of LLT and TFLT increments is theoretically possible. The columns 434 in the TFLT palette 430 contain a series of LLTs in ascending order of thickness from left to right. The rows 432 in the TFLT palette 430 contain a series of ALTs in ascending order of thickness from top to bottom. The sixteen (16) LLT increments provided in the columns 434 in the TFLT palette 430 are 25, 50, 75, 80, 90, 100, 113, 125, 138, 150, 163, 175, 180, 190, 200, and 225 nanometers (nm). The nine (9) ALT increments provided in the rows 432 in the TFLT palette 430 are 0.25, 0.5, 0. 75, 1.0, 1.25, 1.5, 1.75, 3.0 and 6.0 µm. As another example, as illustrated in <figref idref="f0038">Figure 37B</figref>, the LLTs in the columns 434' in the TFLT palette 430' are provided in increments of 10 nm between 0 nm and 160 nm. The nine (9) ALT increments provided in the rows 432' in the TFLT palette 430 are 0.3, 0.5, .0 8, 1.0, 1.3, 1.5, 1.8, 2.0 and 5.0 µm.

0108As part of a per pixel LLT analysis 344 provided in the post-processing system 262 in <figref idref="f0036">Figure 36</figref>, for each pixel in each of the pre-processed resulting images of the area or region of interest in the tear film, a closest match determination is made between the RGB color of the pixel to the nearest RGB color in the TFLT palette 430 (block 345 in <figref idref="f0036">Figure 36</figref>). The ALTs and LLTs for that pixel are determined by the corresponding ALT thickness in the y-axis of the TFLT palette 430, and the corresponding LLT thickness in the x-axis of the TFLT palette 430. As illustrated in <figref idref="f0037">Figure 37</figref>, the TFLT palette 430 colors are actually represented by RGB values. The pixels in each of the pre-processed resulting images of the tear film are also converted and stored as RGB values, although any other color representation can be used as desired, as long as the palette and the image pixel data use the same representational color space. <figref idref="f0039">Figure 38</figref> illustrates the TFLT palette 430 in color pattern form with normalization applied to each red-green-blue (RGB) color value individually. Normalizing a TFLT palette is optional. The TFLT palette 430 in <figref idref="f0039">Figure 38</figref> is displayed using brightness control (i.e., normalization, as previously described) and without the RGB values included, which may be more visually pleasing to a clinician if displayed on the display 174. The GUI utility 280 allows selection of different palettes by selecting a file in the palette file drop down 402, as illustrated in <figref idref="f0025">Figure 27</figref>, each palette being specific to the choice of 2-wave vs. 3-wave mode, the chosen source's spectrum, and the chosen camera's RGB spectral responses. To determine the closest pixel color in the TFLT palette 430, a Euclidean distance color difference equation is employed to calculate the distance in color between the RGB value of a pixel from the pre-processed resulting image of the patient's tear film and RGB values in the TFLT palette 430 as follows below, although the present invention is not so limited: <maths id="math0002" num=""><math display="block"><mi>Diff</mi><mo>.</mo><mo>=</mo><mo>√</mo><mfenced separators=""><msup><mfenced separators=""><mi>Rpixel</mi><mo>−</mo><mi>Rpalette</mi></mfenced><mrow><mn>2</mn><mo>+</mo></mrow></msup><msup><mfenced separators=""><mi>Gpixel</mi><mo>−</mo><mi>Gpalette</mi></mfenced><mrow><mn>2</mn><mo>+</mo></mrow></msup><msup><mfenced separators=""><mi>Bpixel</mi><mo>−</mo><mi>Bpalette</mi></mfenced><mn>2</mn></msup></mfenced></math><img file="EP2413699B1_D0002.tif" /></maths>

0109Thus, the color difference is calculated for all palette entries in the TFLT palette 430. The corresponding LLT and ALT values are determined from the color hue in the TFLT palette 430 having the least difference from each pixel in each frame of the pre-processed resulting images of the tear film. The results can be stored in RAM 258 or any other convenient storage medium. To prevent pixels without a close match to a color in the TFLT palette 430 from being included in a processed result of LLT and ALT, a setting can be made to discard pixels from the results if the distance between the color of a given pixel is not within the entered acceptable distance of a color value in the TFLT palette 430 (block 346 in <figref idref="f0036">Figure 36</figref>). The GUI utility 280 in <figref idref="f0025">Figure 27</figref> illustrates this setting such as would be the case if made available to a technician or clinician. A distance range input box 408 is provided to allow the maximum distance value to be provided for a pixel in a tear film image to be included in LLT and ALT results. Alternatively, all pixels can be included in the LLT and ALT results by selecting the ignore distance selection box 410 in the GUI utility 280 of <figref idref="f0025">Figure 27</figref>.

0110Each LLT and ALT determined for each pixel from a comparison in the TFLT palette 430 via the closest matching color that is within a given distance (if that post-processing setting 266 is set) or for all LLT and ALT determined values are then used to build a TFLT histogram. The TFLT histogram is used to determine a weighted average of the LLT and ALT values for each pixel in the resulting image(s) of the patient's tear film to provide an overall estimate of the patient's LLT and ALT. <figref idref="f0040">Figure 39</figref> illustrates an example of such a TFLT histogram 460. This TFLT histogram 440 may be displayed as a result of the shown LLT histogram selection box 400 being selected in the GUI utility 280 of <figref idref="f0025">Figure 27</figref>. As illustrated therein, for each pixel within an acceptable distance, the TFLT histogram 440 is built in a stacked fashion with determined ALT values 444 stacked for each determined LLT value 442 (block 349 in <figref idref="f0036">Figure 36</figref>). Thus, the TFLT histogram 440 represents LLT and ALT values for each pixel. A horizontal line separates each stacked ALT value 444 within each LLT bar.

0111One convenient way to determine the final LLT and ALT estimates is with a simple weighted average of the LLT and ALT values 442, 444 in the TFLT histogram 440. In the example of the TFLT histogram 440 in <figref idref="f0040">Figure 39</figref>, the average LLT value 446 was determined to be 90.9 nm. The number of samples 448 (i.e., pixels) included in the TFLT histogram 440 was 31,119. The frame number 450 indicates which frame of the resulting video image is being processed, since the TFLT histogram 440 represents a single frame result, or the first of a frame pair in the case of background subtraction. The maximum distance 452 between the color of any given pixel among the 31,119 pixels and a color in the TFLT palette 430 was 19.9, 20 may have been the set limit (Maximum Acceptable Palette Distance) for inclusion of any matches. The average distance 454 between the color of each of the 31,119 pixels and its matching color in the TFLT palette 430 was 7.8. The maximum distance 452 and average distance 454 values provide an indication of how well the color values of the pixels in the interference signal of the specularly reflected light from the patient's tear film match the color values in the TFLT palette 430. The smaller the distance, the closer the matches. The TFLT histogram 440 can be displayed on the display 174 to allow a clinician to review this information graphically as well as numerically. If either the maximum distance 452 or average distance 454 values are too high, this may be an indication that the measured LLT and ALT values may be inaccurate, or that the image normalization is not of the correct value. Further imaging of the patient's eye and tear film, or system recalibration can be performed to attempt to improve the results. Also, a histogram 456 of the LLT distances 458 between the pixels and the colors in the TFLT palette 430 can be displayed as illustrated in <figref idref="f0041">Figure 40</figref> to show the distribution of the distance differences to further assist a clinician in judgment of the results.

0112Other results can be displayed on the display 174 of the OSI device 170 that may be used by a physician or technician to judge the LLT and/or ALT measurement results. For example, <figref idref="f0042">Figure 41</figref> illustrates a threshold window 424 illustrating a (inverse) threshold mask 426 that was used during pre-processing of the tear film images. In this example, the threshold window 424 was generated as a result of the show threshold window selection box 382 being selected in the GUI utility 280 of <figref idref="f0025">Figure 27</figref>. This may be used by a clinician to humanly evaluate whether the threshold mask looks abnormal. If so, this may have caused the LLT and ALT estimates to be inaccurate and may cause the clinician to discard the results and image the patient's tear film again. The maximum distance between the color of any given pixel among the 31,119 pixels and a color in the palette 430 was 19.9 in this example.

0113<figref idref="f0043">Figure 42</figref> illustrates another histogram that may be displayed on the display 174 and may be useful to a clinician. As illustrated therein, a three-dimensional (3D) histogram plot 460 is illustrated. The clinician can choose whether the OSI device 170 displays this histogram plot 460 by selecting the 3D plot selection box 416 in the GUI utility 280 of <figref idref="f0025">Figure 27</figref>, as an example, or the OSI device 170 may automatically display the histogram plot 460. The 3D histogram plot 460 is simply another way to graphically display the fit of the processed pixels from the pre-processed images of the tear film to the TFLT palette 430. The plane defined by the LLT 462 and ALT 464 axes represents the TFLT palette 430. The axis labeled "Samples" 466 is the number of pixels that match a particular color in the TFLT palette 430.

0114<figref idref="f0044">Figure 43</figref> illustrates a result image 428 of the specularly reflected light from a patient's tear film. However, the actual pixel value for a given area on the tear film is replaced with the determined closest matching color value representation in the TFLT palette 430 to a given pixel for that pixel location in the resulting image of the patient's tear film (block 347 in <figref idref="f0036">Figure 36</figref>). This setting can be selected, for example, in the GUI utility 280 of <figref idref="f0025">Figure 27</figref>. Therein, a "replace resulting image..." selection box 412 is provided to allow a clinician to choose this option. Visually displaying interference interactions representing the closest matching color value to the interference interactions in the interference signal of the specularly reflected light from a patient's tear film in this manner may be helpful to determine how closely the tear film interference model matches the actual color value representing the resulting image (or pixels in the image).

0115Ambiguities can arise when calculating the nearest distance between an RGB value of a pixel from a tear film image and RGB values in a TFLT palette, such as TFLT palettes 430 and 430' in <figref idref="f0037">Figures 37A</figref> and <figref idref="f0038">37B</figref> as examples. This is because when the theoretical LLT of the TFLT palette is plotted in RGB space for a given ALT in three-dimensional (3D) space, the TFLT palette 469 is a locus that resembles a pretzel like curve, as illustrated with a 2-D representation in the exemplary TFLT palette locus 470 in <figref idref="f0045">Figure 44</figref>. Ambiguities can arise when a tear film image RGB pixel value has close matches to the TFLT palette locus 470 at significantly different LLT levels. For example, as illustrated in the TFLT palette locus 470 in <figref idref="f0045">Figure 44</figref>, there are three (3) areas of close intersection 472, 474, 476 between RGB values in the TFLT palette locus 470 even though these areas of close intersection 472, 474, 476 represent substantially different LLTs on the TFLT palette locus 470. This is due to the cyclical phenomenon caused by increasing orders of optical wave interference, and in particular, first order versus second order interference for the LLT range in the tear films. Thus, if an RGB value of a tear film image pixel is sufficiently close to two different LLT points in the TFLT palette locus 470, the closest RGB match may be difficult to match. The closest RGB match may be to an incorrect LLT in the TFLT palette locus 470 due to error in the camera and translation of received light to RGB values. Thus, it may be desired to provide further processing when determining the closest RGB value in the TFLT palette locus 470 to RGB values of tear film image pixel values when measuring TFLT.

0116In this regard, there are several possibilities that can be employed to avoid ambiguous RGB matches in a TFLT palette. For example, the maximum LLT values in a TFLT palette may be limited. For example, the TFLT palette locus 470 in <figref idref="f0045">Figure 44</figref> includes LLTs between 10 nm and 300 nm. If the TFLT palette locus 470 was limited in LLT range, such as 240 nm as illustrated in the TFLT palette locus 478 in <figref idref="f0046">Figure 45</figref>, two areas of close intersection 474 and 476 in the TFLT palette 469 in <figref idref="f0045">Figure 44</figref> are avoided in the TFLT palette 469 of <figref idref="f0046">Figure 45</figref>. This restriction of the LLT ranges may be acceptable based on clinical experience since most patients do not exhibit tear film colors above the 240 nm range and dry eye symptoms are more problematic at thinner LLTs. In this scenario, the limited TFLT palette 469 of <figref idref="f0046">Figure 45</figref> would be used as the TFLT palette in the post-processing system 262 in <figref idref="f0036">Figure 36</figref>, as an example.

0117Even by eliminating two areas of close intersection 474, 476 in the TFLT palette 469, as illustrated in <figref idref="f0046">Figure 45</figref>, the area of close intersection 472 still remains in the TFLT palette locus 478. In this embodiment, the area of close intersection 472 is for LLT values near 20 nm versus 180 nm. In these regions, the maximum distance allowed for a valid RGB match is restricted to a value of about half the distance of the TFLT palette's 469 nearing ambiguity distance. In this regard, RGB values for tear film pixels with match distances exceeding the specified values can be further excluded from the TFLT calculation to avoid tear film pixels having ambiguous corresponding LLT values for a given RGB value to avoid error in TFLT measurement as a result.

0118In this regard, <figref idref="f0047">Figure 46</figref> illustrates the TFLT palette locus 478 in <figref idref="f0046">Figure 45</figref>, but with a circle of radius R swept along the path of the TFLT palette locus 478 in a cylinder or pipe 480 of radius R. Radius R is the acceptable distance to palette (ADP), which can be configured in the control system 240. When visualized as a swept volume inside the cylinder or pipe 480, RGB values of tear film image pixels that fall within those intersecting volumes may be considered ambiguous and thus not used in calculating TFLT, including the average TFLT. The smaller the ADP is set, the more poorly matching tear film image pixels that may be excluded in TFLT measurement, but less pixels are available for use in calculation of TFLT. The larger the ADP is set, the less tear film image pixels that may be excluded in TFLT measurement, but it is more possible that incorrect LLTs are included in the TFLT measurement. The ADP can be set to any value desired. Thus, the ADP acts effectively as a filter to filter out RGB values for tear film images that are deemed a poor match and those that may be ambiguous according to the ADP setting. This filtering can be included in the post-processing system 262 in <figref idref="f0036">Figure 36</figref>, as an example, and in step 346 therein, as an example.

<u>Graphical User Interface (GUI)</u>

0119In order to operate the OSI device 170, a user interface program may be provided in the user interface system 278 (see <figref idref="f0022">Figure 25A</figref>) that drives various graphical user interface (GUI) screens on the display 174 of the OSI device 170 in addition to the GUI utility 280 of <figref idref="f0025">Figure 27</figref> to allow access to the OSI device 170. Some examples of control and accesses have been previously described above. Examples of these GUI screens from this GUI are illustrated in <figref idref="f0045 f0046 f0047 f0048 f0049">Figures 44-48</figref> and described below. The GUI screens allow access to the control system 240 in the OSI device 170 and to features provided therein. As illustrated in <figref idref="f0048">Figure 47</figref>, a login GUI screen 520 is illustrated. The login GUI screen 520 may be provided in the form of a GUI window 521 that is initiated when a program is executed. The login GUI screen 520 allows a clinician or other user to log into the OSI device 170. The OSI device 170 may have protected access such that one must have an authorized user name and password to gain access. This may be provided to comply with medical records and privacy protection laws. As illustrated therein, a user can enter their user name in a user name text box 522 and a corresponding password in the password text box 524. A touch or virtual keyboard 526 may be provided to allow alphanumeric entry. To gain access to help or to log out, the user can select the help and log out tabs 528, 530, which may remain resident and available on any of the GUI screens. After the user is ready to login, the user can select the submit button 532. The user name and password entered in the user name text box 522 and the password text box 524 are verified against permissible users in a user database stored in the disk memory 268 in the OSI device 170 (see <figref idref="f0022">Figure 25A</figref>).

0120If a user successfully logs into the OSI device 170, a patient GUI screen 534 appears on the display 174 with the patient records tab 531 selected, as illustrated in <figref idref="f0049">Figure 48</figref>. The patient GUI screen 534 allows a user to either create a new patient or to access an existing patient. A new patient or patient search information can be entered into any of the various patient text boxes 536 that correspond to patient fields in a patient database. Again, the information can be entered through the virtual keyboard 526, facilitated with a mouse pointing device (not shown), a joystick, or with a touch screen covering on the display 174. These include a patient ID text box 538, patient last name text box 540, patient middle initial text box 542, a patient first name text box 544, and a date of birth text box 546. This data can be entered for a new patient, or used to search a patient database on the disk memory 268 (see <figref idref="f0022">Figure 25A</figref>) to access an existing patient's records. The OSI device 170 may contain disk memory 268 with enough storage capability to store information and tear film images regarding a number of patients. Further, the OSI device 170 may be configured to store patient information outside of the OSI device 170 on a separate local memory storage device or remotely. If the patient data added in the patient text boxes 536 is for a new patient, the user can select the add new patient button 552 to add the new patient to the patient database. The patients in the patient database can also be reviewed in a scroll box 548. A scroll control 550 allows up and down scrolling of the patient database records. The patient database records are shown as being sorted by last name, but may be sortable by any of the patient fields in the patient database.

0121If a patient is selected in the scroll box 548, which may be an existing or just newly added patient, as illustrated in the GUI screen 560 in <figref idref="f0050">Figure 49</figref>, the user is provided with an option to either capture new tear film images of the selected patient or to view past images, if past tear film images are stored for the selected patient on disk memory 268. In this regard, the selected patient is highlighted 562 in the patient scroll box 548, and a select patient action pop-up box 564 is displayed. The user can either select the capture new images button 566 or the view past images button 568. If the capture new images button 566 is selected, the capture images GUI 570 is displayed to the user under the capture images tab 571 on the display 174, which is illustrated in <figref idref="f0051">Figure 50</figref>. As illustrated therein, a patient eye image viewing area 572 is provided, which is providing images of the patient's eye and tear film obtained by the video camera 198 in the OSI device 170. In this example, the image is of an overlay of the subtracted first and second tiled pattern images of the patient's tear film onto the raw image of the patient's eye and tear film, as previously discussed. The focus of the image can be adjusted via a focus control 574. The brightness level of the image in the viewing area 572 is controlled via a brightness control 576. The user can control the position of the video camera 198 to align the camera lens with the tear film of interest whether the lens is aligned with the patient's left or right eye via an eye selection control 578. Each frame of the patient's eye captured by the video camera 198 can be stepped via a stepping control 580. Optionally, or in addition, a joystick may be provided in the OSI device 170 to allow control of the video camera 198.

0122The stored images of the patient's eye and tear film can also be accessed from a patient history database stored in disk memory 268. <figref idref="f0052">Figure 51</figref> illustrates a patient history GUI screen 582 that shows a pop-up window 584 showing historical entries for a given patient. For each tear film imaging, a time and date stamp 585 is provided. The images of a patient's left and right eye can be shown in thumbnail views 586, 588 for ease in selection by a user. The stored images can be scrolled up and down in the pop-up window 584 via a step scroll bar 590. Label names in tag boxes 592 can also be associated with the images. Once a desired image is selected for display, the user can select the image to display the image in larger view in the capture images GUI 570 in <figref idref="f0051">Figure 50</figref>. Further, two tear film images of a patient can be simultaneously displayed from any current or prior examinations for a single patient, as illustrated in <figref idref="f0053">Figure 52</figref>.

0123As illustrated in <figref idref="f0053">Figure 52</figref>, a view images GUI screen 600 is shown, wherein a user has selected a view images tab 601 to display images of a patient's ocular tear film. In this view images GUI screen 600, both images of the patient's left eye 602 and right eye 604 are illustrated side by side. In this example, the images 602, 604 are overlays of the subtracted first and second tiled pattern images of the patients tear film onto the raw image of the patient's tear eye and tear film, as previously discussed. Scroll buttons 606, 608 can be selected to move a desired image among the video of images of the patient's eye for display in the view images GUI screen 600. Further, step and play controls 610, 612 allow the user to control playing a stored video of the patient's tear film images and stepping through the patient's tear film images one at a time, if desired. The user can also select an open patient history tab 614 to review information stored regarding the patient's history, which may aid in analysis and determining whether the patient's tear film has improved or degraded. A toggle button 615 can be selected by the user to switch the images 602, 604 from the overlay view to just the images 620, 622, of the resulting tiled interference interactions of specularly reflected light from the patient's tear films, as illustrated in <figref idref="f0054">Figure 53</figref>. As illustrated in <figref idref="f0054">Figure 53</figref>, only the resulting interference interactions from the patient's tear film are illustrated. The user may select this option if it is desired to concentrate the visual examination of the patient's tear film solely to the interference interactions.

0124Many modifications and other embodiments of the invention set forth herein will come to mind to one skilled in the art to which the invention pertains having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. These modifications include, but are not limited to, the type of light source or illuminator, the number of tiling groups and modes, the arrangement of tile groups, the type of imaging device, image device settings, the relationship between the illuminator and an imaging device, the control system, the type of tear film interference model, and the type of electronics or software employed therein, the display, the data storage associated with the OSI device for storing information, which may also be stored separately in a local or remotely located remote server or database from the OSI device, any input or output devices, settings, including pre-processing and post-processing settings. Note that subtracting the second image from the first image as disclosed herein includes combining the first and second images, wherein like signals present in the first and second images are cancelled when combined. Further, the present invention is not limited to illumination of any particular area on the patient's tear film or use of any particular color value representation scheme.

0125Therefore, it is to be understood that the invention is not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. It is intended that the present invention cover the modifications and variations of this invention provided they come within the scope of the appended claims.