Dental extraoral x-ray imaging system and method

Background of the Invention

0001This invention relates to X-ray imaging devices and in particular, devices for dental extraoral X-ray imaging. Firstly, dental panoramic x-ray imaging is a well-known dental radiographic procedure. Its purpose is to produce an x-ray image of the entire jaw for diagnosis as opposed to a partial image such as obtained by intra-oral x-ray imaging. The panoramic image x-ray is particularly useful, for example, when the dentist is planning implants or surgical operations or for orthodontics procedures. A sample dental panoramic x-ray image is shown in <figref idref="f0001">FIG. 1</figref>. The invention relates also to other extra-oral dental x-ray imaging systems such as cone beam dental computed tomography ("CT") system for 3-D tomosynthetic volumetric reconstruction and transverse slicing. These systems are all useful and needed in dental applications depending on the dentists specialization. For example regular panoramic imaging is used mostly from general purpose orthodontists, while 3-D imaging and transverse slicing may be used more often by implantologists.

0002Dental panoramic x-ray imaging units, cone beam units and transverse slicing units, a.k.a. orthopantomographs ("OPGs") or dental CT's, are available from manufacturers including, among others, Instrumentarium, Sirona, Gendex, Planmeca, Schick Technologies, Morita, Yoshida, Asahi, Vatech and others. The units come both with analog film were suitable and with digital sensors (in case of cone beam CT), but otherwise the differences between models of different manufacturers are minor. In all products currently available in the market, digital OPGs are utilizing sensors based on CCDs coupled to a phosphor or scintillator (a material that emits light in response to charged particles in a manner such that when charged particles interact with the scintillator, electrons in the atoms in the scintillator become excited. When the atoms return to ground state, their electrons emit photons) and operating in a Time Delay Integration Mode (TDI), or flat panels utilizing a-Si (amorphous silicon) TFT arrays (Thin Film Transistor) with again a scintillator on the top. Both CCD's and TFT flat panels used convert x-rays to light and then light is converted to an electronic signal inside the CCD or TFT.

0003It should be mentioned from the start that this equipment and especially the cone beam CT and transverse slicing equipment which produce multiple frames are particularly expensive being in the range of 200kUSD to 400kUSD (retail price). Despite the expense, these systems are not fast enough for continuous exposure. Consequently, real time viewing is not possible. Regular digital OPG's are in the range of 40kUSD while "high end" digital OPGs are in the range of 50kUSD-70kUSD. Even so, a fully equipped dental office needs to have several types of so called "intra-oral" sensors, totaling another 15kUSD-25kUSD to complete the range of functionalities needed to cover general maxillofacial examinations, fillings and cavities, orthodontics, implantology and surgery. As can be appreciated, this is a burden that probably only large clinics can afford.

0004Referring now to <figref idref="f0001">FIG. 2</figref>, an OPG is made up of four functional units, namely, an x-ray generator, an imaging device, a mechanical manipulator and a user control panel.

0005The purpose of the x-ray generator is to create the x-rays that penetrate the head of the patient and arrive at the imaging device. The x-ray generator or source is able to generate x-rays with different spectra by varying the high voltage level and with differing intensity by varying the current.

0006The purpose of the imaging device is to detect and convert the incident x-rays into an image. This process used to accomplish this purpose can be either through absorption, by traditional film or digital two-stage indirect conversion (which, in all current commercially available systems, is accomplished using a CCD with a scintillator). Linear arrays of CCD's are used in OPG's but flat panels based on TFT or image intensifiers are used in cone beam dental CT. The cost of square or rectangular TFT arrays (10cmx10cm or larger) is in the range of 15kUSD-30kUSD even for moderate volumes.

0007In contrast the assignee of the current invention is the pioneer of digital single-stage conversion CdTe-CMOS or CdZnTe-CMOS sensors. Sensors of this type are disclosed, for example, in <patcit id="pcit0001" dnum="WO2004055550A"><text>WO2004055550</text></patcit>, and <patcit id="pcit0002" dnum="EP1520300A"><text>EP1520300</text></patcit>.

0008The purpose of the mechanical manipulator is to displace both the imaging device and the x-ray generator in such a way that a proper panoramic image of the plane-of-interest is formed. A user control panel or the user interface, is used to control different settings of the OPG or to initiate and control an x-ray exposure.

0009A typical cone beam dental CT system does not differ in any substantial way from the OPG system, except that the x-ray beam is cone shaped rather than fan shaped. Additionally the cone beam systems require that the x-ray scan be performed for longer times and in steps (ie not continuously) because Image Intesifiers (IIs) or TFT panels are too slow and lack sensitivity.

<i>The dental panoramic, dental transversal and dental 3-D x-ray imaging process</i>

0010A dental panoramic x-ray image is captured during a process in which both the x-ray generator and the imaging device move around the patient's head according to a predetermined geometric path and speed profile. The movement is synchronized in such a way that an image of the pre-determined layer of interest is formed according to the predetermined geometry and speed profile. Because of the shape of the human jaw, this layer is a non-planar structure. It in fact varies with the morphology of each individual's jaw but we will not discuss here an embodiment, which takes inputs measured from each patient's jaw and adjusts the synchronization to follow that unique path for each patient.

0011To simplify the procedure while still maintaining high resolution, a standard shape that is applied to all human males, females, and children of certain ages, is used. The exact shape of the layer-of-interest depends therefore on the dental procedure in question; the predetermined geometric path of the source and detector (optionally varying depending on the patient type) and the predetermined speed profile. Generally, the shape is as illustrated in <figref idref="f0002">FIG. 3</figref>. The layer can usually be adjusted by selecting a different pre-determined, preset program in the OPG by changing the path of movement and/or also the speed profile. Different programs can alter the general parameters of the profile to match the patient (again, e.g., whether child or adult) or to only image a part of the full profile (i.e., front teeth, left/right side etc.). But in each case when a new panoramic layer is needed a new exposure needs to be taken, which means additional radiation to the patient.

0012The movement of the x-ray generator and the imaging device is traditionally synchronized so that the imaging device surface normal is perpendicular to the layer-of-interest. In this way the formed image is distorted as little as possible. A disadvantage of this approach is that the movement trajectory is quite complex. To achieve this motion, multiple motors are required (i.e. degrees of freedom) which also complicates the control electronics and algorithms, thus leading to higher cost. There are some imaging modalities in which the direction of radiation is intentionally not perpendicular to the surface normal, but the same drawbacks and advantages apply.

0013In addition, one of the most severe issues experienced today in clinical applications of dental panoramic imaging is that the patient (object) does or cannot necessarily remain motionless for the whole duration of the scan (typically lasting 5 to 30 seconds). Even a small misalignment of the patient would require the part of the desired layer be reconstructed blurred or out of focus.

0014Referring now to <figref idref="f0002">FIG. 4</figref>, in addition to panoramic images, a dentist may wish to create a transverse slice image of the patient's jaw. In transversal imaging, the layer-of-interest is perpendicular to the panoramic layer.

0015The existing commercially available, extra-oral (including, for example, panoramic and transversal) imaging solutions are either based on slot-shaped (i.e. with an aspect ratio --length "m" divided by width "n" -- of m/n=5 or more) time-delayed integration CCD sensors (not producing multiple frames) or large-area 2D detectors with a computed tomography system backend where m/n is substantially equal to 1 (which, however, do produce multiple frames). The large-area 2D detectors are most often TFT panels and are particularly expensive because of the m/n ≈ 1 aspect ratio (approximately equal to one).

0016The slot-shaped detectors which use a CCD coupled with a scintillator apply the time-delayed integration (TDI) principle to form the image of the layer of interest. Time-Delay Integration is a method of synchronizing the shifting of the image signal captured in the pixel with the movement of the object image across the face of the CCD. This permits integration of more signal, increasing sensitivity, reducing noise and reducing image blur. According to this method, the integrated charges are clocked inside the detector logic (CCD) in the direction of the movement. Thus, at a given integration period <i>t_i</i>, the charge for a fixed object volume <i>v</i> is integrated to a pixel <i>p_n</i>. The object is moved so that the image of the plane-of-interest is moved (taking the magnification factor into account) exactly a pixel's width. After the integration period, the charges are transferred in such a way that if the image of the volume <i>v</i> is projected to pixel <i>p</i>_{<i>i</i>-<i>l</i>}, the charged from pixel <i>p_i</i> is transferred to pixel <i>p</i>_{<i>i</i>-<i>l</i>}. The last pixel value in the row, which has no neighbor to which to transfer the charge, is read out and stored in the final image. In this way, the apparent integration time of an image pixel is the integration period multiplied by the width of the imaging device in pixels.

0017With the TDI principle, the clocking of the charges must be synchronized so that the apparent speed of the layer-of-interest in the imaging devices active in an integration period must be exactly the width of the pixel. If the speed is not matched, the image will appear blurred. A 2D flat image is formed from a single scan using an imaging device operating in the TDI mode. <i>Multiple panoramic layers, transverse slicing or 3D imaging is not possible</i> because only a single projection is saved.

0018In a dental cone beam computed tomography system (3-D imaging), multiple, non-TDI exposures are taken with a 2D area detector where m substantially equals n (i.e. within 20%). The movement is stopped before the exposure and the x-ray source is only active during this stationary period. The movement is continued after the exposure. In this manner, the movement doesn't have to be synchronized. Such systems require a higher dose to be administered to the patient and also longer examination times. The final image is formed as a layer calculated from the volumetric (3D) dataset constructed from the individual exposures or projections. The clear advantage of this method is that a full 3D volumetric dataset will be available after the procedure. However, with current solutions, the resolution of panoramic layer calculated from the 3D data is low compared to dedicated panoramic imaging systems (OPGs). Furthermore, the dose levels are much higher and maybe equally importantly the cost of such available systems is as mentioned in the range of 200kUSD-400kUSD because of the use of large and expensive flat TFT panels, expensive Image Intensifiers or simply because dental OEMs market these "advanced" systems at a price premium.

0019<patcit id="pcit0003" dnum="US6496557B"><text>US 6,496,557</text></patcit>, entitled <i>"Two Dimensional Slot X-ray Bone Densitometry, Radiography and Tomography"</i>, describes a process in which multiple layers are formed by a so-called shift-and-add algorithm. The process described includes a system in which the motion of the imaging device is either linear or includes a rotational component around the focal point of the x-ray source. Unfortunately, such a system cannot be used in the field of dental x-ray imaging where the layer(s) of interest run around or across the human jaw. Although such an approach may be useful in bone densitometry and in some other applications, it is, in practice, impossible to apply to dental panoramic or transversal imaging due to the fact that x-rays run essentially parallel to rather than across the layer-of-interest, should there be linear movement or rotation around the focal point. Secondly the process disclosed in <patcit id="pcit0004" dnum="US6496557B"><text>US 6,496,557</text></patcit> would have another serious limitation if an attempt were to be made to apply it in the field of dental imaging. That limitation is the most serious issue in panoramic imaging and it results in partial image blurring due to misalignment of the patient or due to patient movement. Thirdly, <patcit id="pcit0005" dnum="US6496557B"><text>US 6,496,557</text></patcit> fails to address the need of a system or a procedure that is able to simultaneously perform both panoramic as well as transverse imaging. Further, this patent fails to disclose a system operable in a dental imaging environment.

0020<patcit id="pcit0006" dnum="US5784429A"><text>US 5,784,429</text></patcit>, entitled <i>"Dental panoramic X-ray imaging apparatus",</i> describes a system in which multiple layers are calculated using plural tomographic images corresponding to plural tomographic planes which are arranged at predetermined intervals along the direction of the X-ray irradiation. A convolution process or a frequency process is conducted on a specific tomographic image by using image information of at least one of the tomographic images, so as to remove blur from the specific tomographic image. This patent describes a means of implementing different layers by using image intensifiers, CCD's or combinations thereof.

0021<patcit id="pcit0007" dnum="WO02052505A"><text>WO02052505 A</text></patcit>, entitled <i>"Radiation Imaging System and Scanning Device"</i>, discloses the use of frame mode CdTe-CMOS detectors for creating tomographic images in several applications including dental panoramic images using a dental x-ray setup. In <patcit id="pcit0008" dnum="WO02052505A"><text>WO02052505 A, Spartiotis et al</text></patcit> disclose that the read-out speed or frame rate must be high enough to allow the detector to move by no more than half a pixel size or preferably even less per a read-out cycle. However, if a system is implemented in such a way, serious limitations can arise such as the need to transfer a disproportionately large amount of data to a computer or the like in real-time. Spartiotis et al. remain silent on the management of data to reconstruct and display an image in real-time, during the exposure. Furthermore Spartiotis et al. say nothing about one of the most serious issues facing dental radiology, that of partial image blurring due to patient misalignment. In such cases, part of the image, i.e. the panoramic layer-of-interest, is blurred while another part is in focus. Spartiotis <i>et al</i>. merely disclose that the frame rate needs be very high to collect frames in such a case in which the detector has moved by less than half a pixel. Unfortunately, in this case, the amount of data produced during the exposure is unnecessarily high while at the same time, no material performance benefit is gained.

0022Additionally Spartiotis <i>et al.</i> do not suggest a means by which one might combine data to gain performance benefits and to correct blurring or to produce transversel image slices or even tomosynthentic 3D images.

0023These documents <patcit id="pcit0009" dnum="US5784429A"><text>US 5784429</text></patcit>, <patcit id="pcit0010" dnum="US6496557B"><text>US 6496557</text></patcit> and <patcit id="pcit0011" dnum="WO02052505A"><text>WO02052505A</text></patcit> disclose techniques only in their nearest accordance, according to the preamble part of independent claims.

0024Devices exist that are capable for example of performing transverse slicing or 3-D reconstructed images but most often they require much longer x-ray scan times and non-continuous scan (ie a step by step scan). The long exposures are needed because the normally used digital imaging devices lack sensitivity and typically can "catch" only 1 out of 3 incoming x-rays. Further, non-continuous step-wise scans are necessary in the prior art because of the slow response and slow readout of the current rectangular or square flat panel TFT arrays. Higher doses and longer step-wise scans create higher risk and discomfort to the patient.

0025What is needed therefore is a system that limits the radiation doses a patient receives while maximizing the data output. What is needed is a system which permits continuous fast, real-time x-ray scans. What is needed is a system and method of combining data from a single exposure to not only reconstruct a panoramic layer of interest but to also be able to correct part of the image that is blurred and furthermore produce transversal image slices and 3D images. Still further, what is needed is a system which minimizes the limitations and deficiencies of prior art thus mitigating or completely eliminating such disadvantages and deficiencies.

Summary of the Invention

0026An extra-oral dental x-ray imaging system is <b>characterized in that</b> what is said in the characterizing part of an independent system claim directed to an extra-oral dental x-ray imaging system.

0027A multi-layer dental panoramic and transversal x-ray imaging system includes an x-ray source; a digital imaging device capable to frame mode output with a high frame rate; a mechanical manipulator having at least one rotation axis located in a position other that the focal point of the x-ray source; means of detecting the camera position in 1D, 2D or 3D depending of the complexity of the trajectory; means of reconstructing the final images out of the stored frames; a real time storage system such as RAM, hard drive or an drive array able to store all the frames captured during an exposure; and a digital processing unit capable of executing the reconstruction algorithm, interconnected in operational arrangement. The system is adapted to produce selectively two or more of the following: 1) a predetermined dental panoramic layer image 2) at least part of a non predetermined dental panoramic layer image 3) transverse slice to a selected part of a dental panoramic layer image 4) 3-D reconstruction of a volume corresponding to some part of a dental panoramic layer, all from image data from a single exposure.

0028In particular, the system combines a fast readout with a memory of comparable speed, i.e., a memory which stores and accesses information substantially in any order and in which all storage locations are substantially equally accessible. Such a memory has sufficient speed for storing the multiple frames concurrently with the exposure in a manner which permits retrieval and display in real time.

0029Preferably such system should employ an x-ray imaging device having an aspect ratio of m/n> 1.2 (one point two) where m is the long dimension and n is the short dimension of the active area of the imaging device.

0030In an advantage, the system uses little no more than the radiation dose used in a regular dental panoramic exposure and performs the x-ray scan to collect such data in a continuous movement.

0031In another advantage, the detector device produces a variety of formats of panoramic images from a high-speed, x-ray digital imaging device.

0032In another advantage, the system and method combines data from a single exposure to not only reconstruct a panoramic layer of interest but to also be able to correct part of the image that is blurred and furthermore produce transversal image slices and 3D images.

Brief Description of the Drawings

0033<ul id="ul0001" list-style="none" compact="compact"><li><figref idref="f0001">FIG. <b>1</b></figref> is an X-ray photograph of a sample dental panoramic x-ray image.</li><li><figref idref="f0001">FIG. <b>2</b></figref> is a schematic view of a dental panoramic x-ray imaging system.</li><li><figref idref="f0002">FIG. <b>3</b></figref> is a schematic view of a panoramic layer-of-interest.</li><li><figref idref="f0002">FIG. <b>4</b></figref> is a schematic view of a transverse slice across a point on a panoramic layer-of-interest.</li><li><figref idref="f0003">FIG. <b>5a</b></figref> is a schematic view of the readout and recording of an image using a prior art TDI-type device which records only a single layer by outputting image lines using the CCD TDI technique during the x-ray scan.</li><li><figref idref="f0003">FIG. <b>5b</b></figref> is a schematic view of the readout in accordance with the current invention and recording of an image using a high-speed x-ray digital imaging device that outputs many independent overlapping frames.</li><li><figref idref="f0003">Fig <b>6a</b></figref> is a photo of an example of a high speed CdTe-CMOS x-ray digital imaging device in accordance with the invention that outputs many independent image frames.</li><li><figref idref="f0004">Fig <b>6b</b></figref> is a schematic representation of a CdTe-CMOS hybrid which is used to build up the imaging device of <figref idref="f0003">Fig <b>6a</b></figref> in accordance with he current invention</li><li><figref idref="f0005">FIG. <b>7</b></figref> is a schematic view of the reconstruction of an image in accordance with the invention using the frames from the high-speed x-ray digital imaging device by means of shift and addition of the individual frames.</li><li><figref idref="f0007">FIG. <b>8a</b></figref><b>-b</b> is a flow chart of the Time-compensated laminographic reconstruction (shift-and-add submethod) of the invention.</li><li><figref idref="f0009">FIG. <b>8c</b></figref> is a schematic diagram of the pixel integration intervals over time.</li><li><figref idref="f0010">FIG. <b>8d</b></figref> is a flow chart of the auto-focusing feature of a panoramic layer in accordance with the invention.</li><li><figref idref="f0011">FIG. <b>8e</b></figref> is a flow chart of the feature of the current invention for calculating a transversal slice at a point on a panoramic layer</li><li><figref idref="f0012">FIG. <b>8f</b></figref> is a flow chart of the feature of the current invention for calculating a 3-D image at point on a panoramic layer using a narrow beam, limited angle 3D reconstruction</li><li><figref idref="f0012">FIG. <b>8h</b></figref> is a figure showing by way of example an original speed profile and a locally modified profile corresponding to the same exposure</li><li><figref idref="f0006">FIG. <b>9</b></figref> is a top, schematic view of an example of a panoramic, transversal layer formed by concatenating rows from multiple panoramic images from the same location.</li></ul>

Detailed Description of the Preferred Embodiment(s)

0034Referring now to <figref idref="f0001">FIGs. <b>1</b> and <b>2</b></figref><b>,</b> in accordance with the current invention, a dental x-ray imaging system 10 produces panoramic images 12 for use in dental diagnosis and treatment. The system 10, using data generated from a single exposure, selectively produces at least two of the following: 1) a predetermined dental panoramic layer image 2) at least part of another dental panoramic layer(s) 3) transverse slice to a selected part of a dental panoramic layer image 4) 3-D reconstruction of a volume corresponding to some part of a dental panoramic layer.

0035The system 10 preferably employs a detector having an aspect ratio of its long "m" to short "n" dimensions in its "active area" with an m/n ratio greater than 1.5 (one point five). The "active area" is the portion of the x-ray imaging device that is sensitive to detecting x-ray flux. This is advantageous because such a detector is much more economical and practical than a full square area detector. Although digital detector prices are dropping, it is still the case that a full field digital imaging device having sufficient resolution is much more expensive than a slot shaped imaging device. Preferably, the invention uses a digital imaging device which is a CdTe (Cadmium Telluride) or CdZnTe (Cadmium Zinc Telluride) based detector, bump-bonded to readout Complementary Metal Oxide Semiconductor (CMOS) Application Specific Integrated Circuit (ASIC). CdTe and CdZnTe detectors, bump-bonded to CMOS, have been introduced by the assignee of the current invention as suitably fast x-ray imaging devices, due to their high density and absorption efficiency, readout speed and resolution.

0036The dental x-ray imaging system 10 in accordance with the current invention includes an x-ray source 16; a digital imaging device 14 capable to frame mode output with sufficient frame rate; a mechanical manipulator 20 having at least one rotation axis located in a position other that the focal point of the x-ray source and means (<figref idref="f0007 f0008 f0009 f0010 f0011 f0012">FIGs. <b>8a-8f</b></figref>) for reconstructing at least two final images out of the same set of stored frames 40 from a single exposure.

0037The digital imaging device has active area dimensions 530 m (long) and 510 n(short) whereas m/n>1.5 thus it has a slot shape and is able to readout the frames 40 at 50fps (frames per second) or more preferably 100fps and even more preferably 300fps or more. The m/n>1.5 geometry helps keep the costs of such a device down and indeed costs can be much less that a full field active area imaging device. This is particularly true where the imaging device is a CdTe-CMOS 560 or CdZnTe-CMOS imaging device.

0038Furthermore, the x-ray source arm 20 and digital imaging device 14 move in a continuous scan movement (not necessarily constant) for the duration of the useful part of the scan, i.e. the part of the scan that x-rays are emitted from the x-ray source 16 as needed to image substantially the whole jaw 19. The duration of the scan and the amount of x-rays emitted are comparable to the radiation dose required in a regular dental panoramic x-ray scan.

0039The final images can be selected from a group of predetermined panoramic layer images 99, other panoramic layers images (or at least part thereof 98), transverse slicing images, or 3-D images.

0040A real time storage system 10 such as RAM, a hard drive or a drive array is able to store all the frames 40 captured during an exposure. A digital processing unit, which can be a typical computer CPU or the like, is capable of executing the reconstruction algorithm, interconnected in operational arrangement. The system 10 selectively produces dental panoramic x-ray images or parts thereof, dental transversal x-ray images and dental tomosynthetic 3D images from a frame stream produced by the high-speed, x-ray digital imaging device.

0041Referring now to <figref idref="f0001">FIGs 1 and 2</figref>, the invention described here is a system 10 which selectively produces dental panoramic x-ray images, dental transversal x-ray images and dental tomosynthetic 3D images (hereafter dental x-ray images 12) from a frame stream produced by a high-speed, x-ray digital imaging device 14.The main components of the system 10 includes an x-ray source 16, an imaging device 14, mechanical manipulator 20 and a computer system 22, an alignment estimator <b>24,</b> reconstruction algorithm (<figref idref="f0007 f0008 f0009 f0010 f0011 f0012">FIGs <b>8a-8f</b></figref>), a storage system <b>30</b> that is capable of temporarily storing all the data in effectively real time and a processing unit such as a personal computer 22.. The x-ray source 16 is an x-ray source such as an x-ray generator or a radionuclide.

0042The imaging device 14 is a digital imaging device capable to frame mode output with a sufficient fast frame rate. In accordance with the current invention, imaging device 14 is of the type shown in the photo in <figref idref="f0003">Fig <b>6a</b></figref> which has been implemented by the assignee of the current invention. The slot shaped active area 520 of the imaging device comprises six CdTe-CMOS or CdZnTe-CMOS hybrids. A hybrid is a "sandwich" between CdTe detector substrate and a readout CMOS. Now referring to <figref idref="f0004">FIG. <b>6b</b></figref>, such a CdTe-CMOS hybrid is shown schematically. Substrate 620 can be CdTe or CdZnTe which absorbs the x-rays with extreme sensitivity and converts them directly to an electronic signal. The readout CMOS 630 is interconnected to substrate 620 by means of low temperature lead free solder bumps 610 in accordance with <patcit id="pcit0012" dnum="US6933505B"><text>US6,933,505</text></patcit> belonging to the assignee of the current invention. The individual hybrids are connected electrically to the mother board 560 by means of wire bonding of pads 640 onto the mother board corresponding pads. Beneath the mother board there is situated a readout or interface board 570 for controlling the mother board and producing a digital video signal which is readout via connector 550. Connector 550 can be of the camera link type protocol which is a commercially available readout protocol. But the readout can also happen through a variety of other readout protocols such as USB 2.0, firewire or gigabit Ethernet. The interface board also contains a power supply connector 540 which provides all the necessary supply voltages to the boards as well as High Voltage ("HV") to the CdTe detectors. A frame mode imaging device in accordance with the invention produces independent image frames 40 by means of outputting sequentially or in random access, pixel values every so often and usually in predetermined time intervals. For example a device of the type shown in <figref idref="f0003">FIG. <b>6a</b></figref> has an active area made of an array of 1500 pixels x 64 pixels. The active area is made of six hybrids and combined provides the active slot 520 which has two main dimensions: m and n where m is the long dimension 530 and n is the short dimension 510. As can be seen m/n>>1. Preferably m/n>1.5 and even more preferably m/n>5 or even 10. In the embodiment disclosed in <figref idref="f0003">FIG. <b>6a</b></figref>, the ratio m/n equals 23.4. The reason for that is because the larger the active the more expensive the detector is. Even with mass produced flat panel TFT arrays the price of an imaging device with m/n= 1 or close to 1 is of the order 20kUSD or even 30kUSD. However with an imaging device of the type described herein one is able to perform the functionalities of regular panoramic equipment as well as at least partly the functionality of the very expensive trasversal slicing and 3-D reconstruction dental equipment. Coming back to the concept of a "frame", it is therefore understood that a frame 40 is a two dimensional spatial representation of pixel values 55 each pixel value corresponding to the output from a physical pixel 53 in the imaging device or in some cases each pixel value can correspond to a combination of output values from the imaging device physical pixels 53. The translation of image device pixel values 53 to frame pixel values 55 can take place inside the imaging device or externally, for example inside the computer 22. An imaging device of the type disclosed herein is able to produce many individual frames 40 during the x-ray scan, with very high sensitivity. In a regular CCD based panoramic examination the CCD works in the Time Delay Integration mode (TDI) which is famous in the field. The output from a CCD imaging device is image lines 110 as depicted schematically in <figref idref="f0003">FIG. <b>5a</b></figref>. For the duration of the scan the CCD outputs image lines 110 and at the end of the scan only one panoramic image layer 12 is reconstructed corresponding to a predetermined layer depending on the mechanical geometry of the panoramic unit and the speed profile and the positioning of the patient. Contrary to that an imaging device of the type described in <figref idref="f0003">FIGs <b>6a</b></figref> and <figref idref="f0004"><b>6b</b></figref> can operate in frame output mode, with sufficient speed and excellent sensitivity to match the speed. The imaging device implemented by the applicant operates at frame rates of 50fps to 300fps or more depending on the mechanical scanning speed of the x-ray source. As shown in <figref idref="f0003">FIG. <b>5b</b></figref> during the x-ray scan the output from the imaging device are frames 40 rather than lines, such frames 40 overlapping during the scan and providing the necessary data redundancy needed to reconstruct not one but several dental panoramic layers, or parts of layers, transverse slicing and even 3-D tomosynthetic image reconstruction of teeth. The imaging device of <figref idref="f0003">FIG. <b>6a</b></figref> can also provide just raw independent pixel values and the frames 40 are reconstructed then at the personal computer by rearranging the pixel values to frames.

0043The mechanical manipulator 20 has at least one rotation axis 34 located in a position other that the focal point 36 of the x-ray source. The alignment estimator 24 detects the camera position in 1D, 2D or 3D depending of the complexity of the trajectory. The panoramic reconstruction process 26 uses an algorithm to reconstruct the final images 12 out of the stored frames 40. The storage system 30 is capable of storing the frames 40 in real time in order to avoid frame loss and is therefore essentially a real time storage system such as RAM 42, or very fast hard drive 44 or an drive array 46 capable of transferring stored data at 10 MB/sec or faster and therefore able to store all the frames 40 captured during an exposure while permitting real time retrieval of such data. In a preferred embodiment, RAM 42 having a 50 MB/sec retrieval rate is for example Kingston 400MHz 2 GB DDR (model number KHX3200AK2/2G). These storage means may be located in the computer or located in the imaging device or at some other location. Storing frames 40 in real time here means with a time delay of not more than few seconds and preferably few milliseconds and even more preferably few microseconds. The processing unit 32 is a digital processing unit such as a personal computer 22, micro controller, FPGA or DSP, (not shown) capable of executing the reconstruction algorithm 26.

0044Referring now to <figref idref="f0002">FIGs. <b>3</b> and <b>4</b></figref>, the invention is a fast x-ray imaging device 14 combined with the reconstruction algorithm 26 to allow formation of multiple dental x-ray images from the frames 40 of a single exposure. A single exposure refers here to a continuous measurement in which the whole layer-of-interest 52 (the layer-of-interest can be a part of the whole dental area) is exposed to radiation one or more times, but in which the apparent movement between two consecutive radiated frames 40 is never less than half the size of a pixel 53 but preferably stays less than the size of the pixel. This results in there being ample overlap 56 between consecutive frames 40, but not at the expense of creating huge data sets that cannot be utilized in real time. This is particularly important in dental imaging where the user (dentist) is expecting to see an image 12 in effectively real-time and during the exposure.

0045"Effectively real-time" as used herein means that the image 12 is displayed in less than 10 seconds after the end of the exposure and more preferably less than 5 seconds and, optimally, the image is displayed effectively simultaneously during the exposure. For this to be achieved, the frame rate or time intervals between consecutive frames 40 must not be shorter than what is needed to allow the physical pixels 53 to move by at least half a pixel size but short enough to have the detector pixels move less than a pixel size. The assignee of the current invention has developed such imaging devices 14 as were described herein and with reference to <figref idref="f0003">FIGs <b>6a</b></figref> and <figref idref="f0004"><b>6b</b></figref>, capable of producing multiple frame data with the above qualities, based on CdTe-CMOS or CdZnTe-CMOS technology as described in <patcit id="pcit0013" dnum="WO2004055550A"><text>WO2004055550</text></patcit>, <patcit id="pcit0014" dnum="EP1520300A"><text>EP1520300</text></patcit>.

0046The system 10 of the invention (at least) rotates around an axis 34 which is positioned between (but not necessarily crossing the volume of x-rays 35 from the source 16 to the imaging device 14) the x-ray source and the imaging device.

0047Furthermore, the frames 40 generated by the CdTe-CMOS or CdZnTe-CMOS imaging devices 14 are temporarily stored in real time in a sufficiently large RAM 42, to allow real processing and display of a dental panoramic image 12 of a layer 52. The processing of the frames 40 to produce such panoramic images 12 includes spatial (i.e. pixel) domain arithmetic.

0048The difference between a normal TDI-type device (not shown) and the disclosed device 10 is illustrated in <figref idref="f0003">FIGs. <b>5b</b></figref> and <figref idref="f0005"><b>7</b></figref>. A normal TDI-type device records only a single layer 12 while the disclosed device 10 records multiple overlapping frames 40 which can be used to calculate multiple layers (both panoramic and transversal) and limited tomosynthetic 3D structures.

0049Referring now to <figref idref="f0003">FIGs. <b>6a</b></figref> and <figref idref="f0004"><b>6b</b></figref>, the imaging device 14 itself has an aperture 520 which is a slot-shape (having one edge 530 much longer that the other edge 510).

0050The imaging device 14 and x-ray generator 16 rotate around head of the patient 72, but the angular coverage can be less that a full circle as compared to a CT system (not shown). The position of the imaging device 14 is either recorded as a one-dimensional position counter which tells the position of the current frame 40 in the final panoramic image 12 or it can record the full 3D position including orientation.

0051The system 10 records the position of the imaging device 14 as a function of time and thus it's possible to reconstruct a full dental image 12 from individual frames 40 stored.

0052Referring again to <figref idref="f0001">FIG. 1</figref>, the simplest reconstruction is the one producing a dental panoramic x-ray image 12. In this case, the position of the current frame 40 is recorded as a coordinate 74 in the final image 12. This coordinate 74 is then used to calculate the shift <b>76</b> required in a shift-and-add algorithm, mentioned above as the reconstruction algorithm 26 used in reconstructing the final image.

0053Referring again to <figref idref="f0008">FIG. <b>8b</b></figref>, the sub-pixel shifting is accomplished by adding the pixels in a frame 40 to two locations in the final image 12 multiplied by suitable weighting factors 96, 100. If the target position 86 is x (non-integer or integer) and the position is increasing in a positive direction, then the pixel value 55 is added to positions floor(x) and ceil(x) where floor(x) refers to the largest integer smaller than x and ceil(x) refers smallest Integer larger that x. The respective weighting coefficients 96 and 100 are x-floor(x) and ceil(x)-x. The weighting factors 96, 100 can be global to a single frame 40 or can vary from pixel to pixel to compensate for any time delays between individual pixels.

0054This is mathematically equivalent to interpolating the frames 40 and final image 12 linearly in the horizontal direction, shifting the frame pixels in the horizontal direction by an integer amount and then down-sampling the frames 40 and the final image to the original size. The sub-pixel shifting can also be implemented using any other interpolation method, for example with bi-linear, bi-cubic or spline interpolation.

0055The sub-pixel shift-and-add algorithm 26 is implemented to eliminate extra jaggedness of diagonal edges compared to an integer shift-and-add algorithm (i.e. only use integer part of the position 86 without any interpolation or weighing).

0056Referring now to <figref idref="f0001">FIG. <b>2</b></figref>, the dental x-ray imaging system 10 includes an x-ray source 16, and an x-ray imaging device 14, and an electro mechanical manipulator 20. The x-ray source 16 exposes x-rays to an object 19 to be imaged. The x-ray imaging device 14 is adapted to produce multiple frames 40 during at least part of the exposure. The mechanical manipulator 20 controls the movement of the x-ray source 16 and the imaging device 14 about at least one rotation axis along a spline, which is a non-circular trajectory. The axis is located somewhere between the x-ray source focal point and the x-ray imaging device. A means (see <figref idref="f0010">FIG. <b>8d</b></figref>) is provided which uses the multiple frames 40 to compose a panoramic image of a layer of the object under observation, the image having a focus depth which is different in at least some part of the panoramic image from the focus depth corresponding to a predetermined panoramic image.

0057Referring to <figref idref="f0011">FIG. <b>8e</b></figref>, in another embodiment, the dental x-ray imaging system 10 of the invention includes means of using the multiple frames 40 to compose a transversal slice with respect to a panoramic image.

0058In a typical exposure, the scan time is 5 to 30 seconds. A typical slot-shaped CdTe-CMOS or CdZnTe-CMOS sensor (<figref idref="f0003">FIG. <b>6a</b></figref>) that the assignee uses in a dental system is 150mm x 6.4mm in size. The pixel size is typically 100um, although smaller pixel sizes are achievable (but considered not necessary in dental extra-oral imaging). With a frame rate of 200 to 300 fps (frames per second), the CdTe-CMOS sensor moves about half a pixel size to less than one pixel size between consecutive frames 40. At that rate, the output frame data 40 can reach more than 750MB for the entire scan which is a large dataset, but still is quite manageable. It is essential therefore that the frame rate not be too high as this would not add anything to the image resolution and would only make data transfer and processing difficult, if not impossible, in real-time.

0059The assignee of the current invention uses a video or frame grabber 125, based on a technology such as (but not limited to) "CAMERALINK"™, installed to a PCI/PCI Express slot 126 of the PC 22 to capture frames 40. The frames 40 are then stored temporarily in RAM 42. This is very important because otherwise it would be very difficult, if not impossible, to store and process 750MB of data and display an image 12 that is reconstructed from such a large data set in real-time. In accordance with the current invention, the use of computer RAM 42 enables temporary storage of data.

0060Referring in particular to <figref idref="f0007">FIG. <b>8a</b></figref> and <figref idref="f0008"><b>8b</b></figref>, the shift-and-add submethod 26 is a fast processing method for reconstructing and displaying a panoramic layer 12 in real-time, and includes the following steps. In a first step 26a the final image 12 is initialized with zero values. In a second step 26b, all the collected frames 40 are then processed one-by-one in the following manner: in a first substep 26b', for each frame pixel in an individual frame, the pixel position 86 is calculated. In a second substep 26b", weighting coefficients 96 and 100 are calculated according to <figref idref="f0008">FIG. <b>8b</b></figref>. In a third substep 26b"', the pixel value multiplied by the weighting factors is added to the pixel value of the final image 12 in locations specified in <figref idref="f0008">FIG. <b>8b</b></figref>.

0061After the display and processing, the data set 132 can be stored on a hard drive 44, CD 45, DVD 48 and other media. The use of RAM 42 to store the data temporarily and the fast reconstruction algorithm 26 mentioned above allows the image 12 to be reconstructed and displayed no later than 10 seconds from the end of the exposure and usually within 5 seconds from the exposure. In fact, the assignee has currently implemented systems (not shown) in which the image 12 is reconstructed from 750MB of single frame data 134 and it is displayed actually in real-time with a minimal delay (needed for causality) during the actual exposure.

0062In another embodiment, a dental x-ray imaging system 10 further includes an x-ray imaging device 14 which produces multiple frames 40 during an exposure with time intervals during which the detector pixels have shifted by at least half a pixel length or more in the direction of scanning.

0063In another embodiment, the system 10 includes a means of using the multiple frames 40 to compose a panoramic image of a layer in real-time. Real time in this context is defined as having a time lag of not more than 10 seconds from the end of exposure and preferably less than 5 seconds and even more preferably almost concurrently with the exposure.

0064In another embodiment, the system 10 includes a memory such as a RAM (random access memory) for storing the multiple frames 40.

0065In another embodiment, the system 10 includes an x-ray imaging device which produces multiple frames 40 during the exposure with time intervals during which the detector pixels have shifted by at least half a pixel length but less than one pixel length in the direction of scanning.

0066In accordance with another embodiment of the invention, all the individual frames 40 are stored temporarily on RAM 42, and therefore it is possible to reconstruct the final image 12 with a modified position profile, after the actual exposure and after the initially displayed panoramic image 12. It is known of course, that the data can also be stored on a hard drive 44, CD 45 or DVD 48 and retrieved later. However, storing all the data on RAM 42 makes processing fast and efficient. There is no need to wait for data to be downloaded, a process that normally takes several seconds. This, however, is not possible with TDI-type imaging devices, because the position synchronization has already been done in the camera hardware in analog domain and thus cannot be modified any more.

0067The shifted amount 76 between consecutive frames 40 is normally determined from the position information gathered during the measurement. This should give a panoramic image 12 with the layer-of-interest 52 in full focus. Unfortunately, due to misalignment of the patient 72, or patient motion, different jaw profile or for many other reasons, there are very often parts in the panoramic image 12 that are out-of-focus or, simply said, blurred.

0068The shift 76 between consecutive frames 40 (i.e. the speed of the movement in the imaging plane 12) determines which layer 99 is well focused. If the movement speed of the camera 14 with respect to the x-ray source 16 is <i>v_camera,</i> the distance between x-ray source focal point 36 and the imaging plane 12 is <i>d_i</i>, and the layer of interest 52 is at distance <i>d₂</i> from the focal point 36, then the required speed <i>v_shift</i> for the shift-and-add algorithm 26 is determined according to equation (1) <maths id="math0001" num="(1)"><math display="block"><msub><mi>v</mi><mi mathvariant="italic">shift</mi></msub><mo>=</mo><mfrac><mrow><mi>d</mi><mn>1</mn></mrow><mrow><mi>d</mi><mn>2</mn></mrow></mfrac><msub><mi>v</mi><mi mathvariant="italic">camera</mi></msub></math><img file="EP2223653B2_D0001.tif" /></maths>

0069Thus, by modifying the shifting amount 76, a different layer 99 can be focused. In theory; any layer 99 can be in focus, but the finite pixel size sets a limitation on how sharply a layer appears. If the shifting speed 76 is more than a pixel 53 per frame, the image resolution is degraded in the process.

0070The shifting amount 76 can be modulated by software either globally or locally to allow either a completely different layer 99 to be displayed or to modify an existing layer to improve the sharpness in out-of-focus regions, which is usually the case in dental panoramic x-ray imaging.

0071The shift speed 76 cannot only be modified on a column-by-column basis. In addition, the speed 76 can vary on a row-by-row basis (i.e. every part of the image 12 might have a different speed). This enables, for example, the bringing into focus of both upper and lower teeth 180 and 182, respectively, if the patient 19 has an incorrect bite (i.e. front and back teeth are not on the same vertical plane). These aspects of the invention have tremendous impact on today's practice of dental radiology because the current invention makes possible risk free exposures, since a blurred part in the image 12 and anywhere in the image can be focused and displayed with the rest of the original panoramic image. Therefore it is no longer necessary to expose the patient 19 a second, much less a third time.

0072In the above manner, the layer-of-interest 52 can be modified by altering the shifting amount 76 in the shift-and-add algorithm 26. In accordance with another embodiment of the current invention, the optimal speed profile 186 may be calculated. Further, a method 190 is provided to automatically calculate the layer-of-best-focus 192 for dental panoramic imaging.

0073This is achieved by calculating many layers 60 or parts 98 of layers and calculating a sharpness measure 1130, in accordance with <figref idref="f0010">Figure 8d</figref>. Based on this sharpness measure 1130, which can either be a measure of contrast, roughness or some other measure of the image sharpness, the sharpest layer 60 is selected. The algorithm 26 can be applied globally to the whole final image 12 or locally to a given region-of-interest 98. Therefore the dentist is able to observe and initial panoramic image 12 and then select a region (part) 98 of the image 12 where blurring may be evident, in which case, the above mentioned algorithm 27 maximizes sharpness of the selected part of the image. The result is a complete image 12 with all parts well in focus.

0074By reversing the direction in which the frames 40 are added to the final image 12, a completely different layer 60 can be reconstructed. By using the same speed profile 186, only reversed a direction, a mirror layer can be reconstructed. This layer is on the opposite side of the patient and in the dental panoramic imaging, this equates to the area of the neck in the center part of the scan. So the invention provides a method to calculate the image of the spine. By re-projecting this image to the normal panoramic layer (by flipping the mirror image and applying suitable rescaling function), an estimate of the blurring caused by the spine is obtained. This estimate can then be subtracted from the panoramic image 12 to decrease the effect of the spine on the image quality. As an end result, the panoramic image 12 is obtained without any potentially distracting superfluous images of the spine structure.

0075In another embodiment, the system 10 is adapted to create panoramic dental images 12 whereas the focus depth of a panoramic image is different in only some portions when compared to the focus depth of a predetermined panoramic image.

0076In still another embodiment, the system 10 is adapted to create panoramic dental images 12 in which the means of using multiple frames 40 comprises sub-pixel shifting and the adding of pixel values from different frames.

0077In another embodiment, the system 10 includes a means for creating a transversal dental image slice 224. This means uses multiple frames 40 to form the transversal image 224 by combining vertical pixel rows referring to the same physical transverse slice from multiple panoramic images.

0078An important embodiment of the current invention is the system 10 that forms a transversal image slice 224 in addition to a panoramic layer 60 using the frames 40 accumulated during the same exposure. In transversal imaging, the layer-of-interest 69 is perpendicular to a panoramic layer 52. Usually this is done as described earlier, but in accordance to the invention, a new and more elegant method 228 is presented.

0079Referring again to <figref idref="f0011">FIG. <b>8e</b></figref>, as described earlier, multiple panoramic layers can be calculated from a single measurement dataset. By intelligently calculating correct speed modulation function, a transversal slice image 224 can be formed by concatenating the corresponding rows from individual panoramic layers in order of increasing depth. The algorithm 228b (<figref idref="f0011">FIG. <b>8e</b></figref>) has the following steps: In a first step, method 228b' is applied to calculate multiple panoramic images, stored in a stack. In a second step 228b", the contribution of the 3D physical structure at different distances is calculated for every panoramic image 12. Every panoramic image 12 contains the image of the 3D physical model not only at the layer-of-interest, but also in front of and in back of it. This can be modeled using a blur function. As the last step 228b"', the blurring is reversed using a deconvolution method and thus the true image at different panoramic layers without crosstalk between neighboring layers is formed. This image is the transversal image 224.

0080The transversal image 224 is formed by concatenating different panoramic layers 60. Contrary to <patcit id="pcit0015" dnum="US6496557B"><text>US 6,496,557</text></patcit>, which describes a system where the imaging plane is perpendicular or roughly perpendicular to the direction of radiation, the current invention discloses a system 10 which provides a layer 60 parallel (or close to parallel) to the direction of radiation.

0081Referring now to <figref idref="f0012">FIG. <b>8f</b></figref>, additionally, an imaging method 229 can be used to provide limited volumetric 3D images. The method is an iterative algorithm, which uses the geometry data 1180 (i.e. 3D location of x-ray source and the detector and the related movement profiles) to form reprojection data 1190 (i.e. estimates of the projected frames 40 based on the current 3D model). Then the software calculates the error 1200 between measured and reprojected frames 40. This error is used to update the current 3D estimate 1220. This operation will be applied to a region-of-interest 1230 specified by the user. Because of the limited number of views, narrow projection rotational angles and narrow projection width, it's very important to use prior data 1210 in the reconstruction. Prior data is used to set restrictions on the 3D model, like smoothess, to produce and unique and meaningful result.

0082The 3D image obtained is not comparable to full CT images in quality, but is sufficient for dental operations.

0083An object of the invention is to produce selectively predetermined dental panoramic image layers, dental panoramic x-ray images of different layers, dental transversal x-ray images and dental x-ray images from a frame stream produced by a high-speed, x-ray digital imaging device.

0084In an advantage, the system's 10 functioning on data from a single exposure spares the patient from multiple exposures and allows for the first time effectively risk free exposures in dental radiology. Such system 10 has unique advantages and offers a breakthrough in the field of dental radiology.

0085In another advantage, the dentist is able with a single system 10 and single exposures to do examinations that previously required at least two if not more systems totaling a multiple in cost and a multiple of risk in patient diagnosis and treatment. Examinations become safer and less expensive for the patients who now only need take one exposure.

0086In another advantage, the system 10 applies little more radiation than the radiation dose used in a regular dental panoramic exposure and should employ x-ray scan mechanics that operate in continuous movement.

0087In another advantage, continuous movement means that the speed of scanning is v>0 (greater than zero) at all times during the scan. The speed need not be constant.

0088In an advantage, CdTe or CdZnTe based digital imaging devices have an excellent sensitivity and absorb 95% of the incoming x-rays at dental energies (ie from 10kV up to 90kV x-ray tube load).

0089In another advantage, the need for non continuous step-wise scans are eliminated, thus reducing the radiation doses and eliminating the longer step-wise scans which create higher risk and discomfort to the patient. Therefore a continuous fast x-ray scan is possible and in fact one does not need separate specialized dental equipment other than the normal panoramic x-ray units.

0090In another advantage, a system 10 and method is provided of combining data from a single exposure to not only reconstruct a panoramic layer of interest but to also be able to correct part of the image that is blurred and furthermore produce transversal image slices and 3D images.

0091In another advantage, the system 10 and method of the invention produce a variety of formats of panoramic images from a high-speed, x-ray digital imaging device.

0092Multiple variations and modifications are possible in the embodiments of the invention described here. Although certain illustrative embodiments of the invention have been shown and described here, a wide range of modifications, changes, and substitutions is contemplated in the foregoing disclosure. In some instances, some features of the present invention may be employed without a corresponding use of the other features. Accordingly, it is appropriate that the foregoing description be construed broadly and understood as being given by way of illustration and example only, the scope of the invention being limited only by the appended claims.

ADDENDUM

Glossary of terms

0093<dl id="dl0001"><dt>CCD</dt><dd>Charge coupled device - an imaging device capable of converting light to electric signals. In x-ray imaging, a CCD is usually coupled with a scintillator.</dd><dt>CT</dt><dd>Computed tomography - a method of calculating full volumetric 3D data from multiple projections covering the full 360 degree rotational circle.</dd><dt>Imaging device</dt><dd>A device with multiple elements (pixels) able to convert x-ray radiation to a digital image</dd><dt>Pixel column</dt><dd>a group of pixels perpendicular to a pixel row, i.e. pixels along a line perpendicular to the direction of movement</dd><dt>Pixel row</dt><dd>A group of pixels in an imaging device in the movement direction. I.e. if the camera is rotating around the vertical axis, a pixel row refers to pixels along a horizontal line</dd><dt>Scintillator</dt><dd>A device able to convert x-ray radiation (for example x-rays and gamma) to light. A scintillator is usually coupled with a CCD to provide a device able to convert radiation to electric signals.</dd><dt>TDI</dt><dd>Time-delayed Integration - a method in which a single scanned 2D image is formed by concatenating multiple 1D line images</dd></dl>