A search tool

ES 2 395 085 T3

DESCRIPTION

A search tool

The present invention relates to a method for selecting the start image at which the rendering of an image stream will begin, and a system for selecting a start image at which the rendering of an image stream will begin. It finds particular but not exclusive use in the field of video recorders such as the Sky + ™ Box, as well as in DVD players and computer-implemented video playback systems. Often times a person watching a television program or movie using a personal video recorder may choose to start viewing the program or movie at a particular starting point other than the beginning of the recording. When playing a dVd of for example a sitcom series, such as Friends a viewer can select an episode in which to start watching the DVD. The recording on the DVD is separated into data packets, each corresponding to a different episode. The viewer can only play the recording from the beginning of each data packet. This restricts the viewer to watching the DVD recording from a limited number of specific starting points.

The viewer can choose to skip the commercials over a personal recording of a television broadcast or broadcast. To find the end of each commercial intermission, the viewer must fast-forward through the recording. You could increase the speed at which the recording (input) plays while an accelerated version of the recording is played. This is a time-consuming way to search for recording. You could fast forward through the recording even at an even faster speed. This is a faster way to move through the recording, but the viewer cannot easily tell when the recording has advanced at the end of the commercials without much overshoot. Once the recording plays back, you will likely have to input or review the correct point in the recording.

Sometimes data packets are separated at fixed time intervals through the recording data stream. Selecting one of these data packets prevents a user from easily identifying his preferred restart point, if his chosen starting point occurs within a data packet. To select their chosen starting point, the viewer would have to fast forward or rewind the image stream from the beginning or end of a data packet.

In EP 0999504, an information processing apparatus, an information processing method, and an information provision means allow a user to check the still images of a moving image to search for the desired time points and intervals. of time desired and in the desired amount. The user operates a mouse or keyboard to enter a time and a time interval in which an image is rendered. Based on the time and the time interval, the CPU generates thumbnail images of the movie clip. The CPU then simultaneously renders the generated thumbnail images of the image, arranged in a sequential mode over time.

In US 2003067479, a hierarchical picture indexing method is provided which extracts representative pictures from a TV program and generates a hierarchical index picture guide from the representative pictures, and therefore an apparatus is provided. The respective images at predetermined times are drawn from a predetermined program. A hierarchical index is assigned to each of the extracted images and the indexes are mapped to the locations of the packages in the transport stream. An image tree that has indexes assigned is constructed as a hierarchically indexed image guide (HIIG). When a stored program is to be played, the location to be viewed on the HIIG is selected, and from the selected location, playback and playback strategies (fast forward and rewind) are made possible so that a user searches for the program easily and starts viewing the program at the desired location.

Document EP 0 619 550 A2 discloses a moving image data search function in which a position is initially represented in which a scroll box on a scroll bar shows a relative storage position in the storage data of a currently rendered frame. It is possible to define the region to search by selecting different parts of the scroll bar and the scroll box and then a region of a screen window represents a predetermined number of reduced frames, these being the representative frames (divide the search region into 10 slices representing a 3x3 mosaic of representative screens, for example). When the user specifies the area to search (selecting representative frames as a new start point and / or end point on the screen) the CPU renders an image around all representative frames to inform the re-specified search area. The CPU then divides the specific area to search, which is the area between the stored screen images, corresponding to the two representative screen images selected on the display screen into 10 portions to determine representative screen images. of a new second state. The representative images of the determined screen are represented on the screen. After this, the procedure is repeated so that the reduction of the search area and the displacement of the representative images on the screen are carried out.

The non-patent article by HONGJIANG ZHANG et al .: CONTENT BASED VIDEO BROWSING TOOLS PROCEEDINGS OF THE SPIE volume 2417, pages 389, 398, discloses a content-based hierarchical search engine.

ES 2 395 085 T3

The aim of this invention is to reduce the aforementioned drawbacks of the prior art.

In a first aspect of the invention there is a method as defined in claim 1.

In this specification, the closest candidate image is referenced. When the closest candidate image is selected, this is not necessarily the candidate image that is closest in time to the start image. It will be appreciated from this specification that when a user selects the last of the candidate images that it recognizes before the start image, that last candidate image is considered the closest to the start image since the user would not wait to recognize recurring candidate images later in the image stream. The closest candidate image phrase should be interpreted accordingly.

This procedure allows a viewer to quickly identify the point at which they choose to start, or restart, a program.

In the step of iteratively re-selecting the reference image, the range of the searched image stream can be reduced in successive iterations. Advantageously, the scope of the search is reduced over each of the successive iterations.

It is preferred that the method includes one or more iterative stages of rendering a plurality of candidate images within a time interval of the image stream near the candidate image selected by a user in response to the previous iteration. This allows the procedure to select the desired start image much more precisely.

When the candidate images are displayed, it is preferable that they are displayed in chronological order, thereby assisting the user in the operation of the procedure. It is more preferable for candidate images to be arranged in the same configuration as the plurality of keys on the user controller, enabling a user to select an image by pressing an associated key. Each of the candidate images can include a corresponding numeric selection. Furthermore, to assist the user, a firm selection of a candidate image can be made which causes the flow of images to be rendered to start from that image, or from an image extremely close to it. It is preferable that the firm selection can be made by holding down a key on the keyboard that corresponds to the selected candidate image for a relatively long duration, for example for more than 0.6 seconds or by pressing the key twice within a short period of time. time, for example within 0.5 seconds. This means that any additional iterations that would normally be taken into account are reversed in favor of the system going directly to reproduce the image stream from that point. Other types of firm selection can be made. For example, the key on the keyboard that corresponds to the selected candidate image could be pressed together with a shift key or an extension key, to distinguish it from a normal keystroke of that key. If the system uses an on-screen selection, then some firm on-screen selection means can be used. For example, if a pointer is used on the screen, the pointer can be moved to the selected candidate image, and a double click can be used to render a firm selection. If a pointer is not used, but candidate images are selected using a focus that moves from image to image on the screen, the appropriate candidate image can be highlighted using focus, and a firm selection is made, for example by a double click. about that selection. When the screen is touch sensitive, a touch of the appropriate candidate image on the screen can be used to make the selection.

In some embodiments, it is advantageous to define the image stream from which candidate images are rendered, including a reference image at the beginning of the image stream. The representation of the candidate images may further comprise the step of representing the time interval of each of the candidate images with respect to the reference image.

The candidate images can be still, or a moving clip of the image stream that gives moving images for each or some of the candidate images. The duration of each clip is preferably no more than two seconds. When the time interval between candidate images is less than the duration of each clip, the candidate images can be still.

The candidate images are preferably spaced by a time interval, and it is further advantageous to reduce the time interval for each of the successive iterations. It is advantageous to reduce the time interval according to a geometric progression.

In some circumstances it may be preferable if the time interval between candidate images is adjustable. This allows the user to increase or decrease the time span between candidate images. Shortening the time frame can help you focus on where you want to start the image flow. Increasing the time frame can allow the user to quickly search through a longer stream of images. The time interval between candidate images can be adjusted by providing a key or keys on the controller, for example the volume up and volume down keys.

Each of the steps of rendering the candidate images may include rendering the reference image as one of the candidate images. The reference image in the preferred embodiment precedes the other candidate images in the image stream. It is also preferable that the procedure also includes

ES 2 395 085 T3 a stage of rendering the start image after a predetermined number of iterations has been executed. This may involve counting the number of iterations executed by the procedure. It may also be advantageous to analyze the time interval between candidate images, or the duration of the image stream, and increase or decrease the number of iterations executed by the procedure accordingly. For example, the number of iterations required to skip 3 or 4 minutes of ads in an image stream is probably less than to find a point within a 5-hour video recording. The processor operating this procedure may include an algorithm that determines the appropriate number of iterations to use.

Advantageously the candidate images are equally spaced along the image stream.

The time interval in the first iteration can be at least 40 seconds. Advantageously, the initial elapsed time between each of the candidate images is longer than a user would normally want to 'feed' forward through the image stream. Preferably, the time interval over the last iteration is 0.625 seconds. As referred to above, it may be advantageous to be able to adjust the time interval between candidate images. A preferred way of doing this is to use a key or keys on the controller to adjust the time interval between candidate images. The keys used could include the volume up and volume down keys. Additionally, the number of iterations can be varied depending on the length of the image stream or the time interval in the first iteration.

There may be nine images represented. There may be nine keys, and the keys could have a regular numeric keyboard configuration, for example with the numbers from 1 to 9. Advantageously the numbers on the keyboard could be represented on the screen, each number with its associated image. The numbers may be arranged on the screen so that they do not obscure the candidate images. As referred to above, the keys can be used to make a hard selection in which the key is held down for a relatively long period of time to begin the flow of images from the candidate image corresponding to the key that was pressed.

The image stream can have an audio component. Advantageously, the method is suitable for application to audiovisual programs, including, but not limited to, television programs, company or home videos.

The image stream may comprise a series of frames. Advantageously the image stream is a recording of a television broadcast, including, but not limited to, television programs, a movie, or a home video. In certain embodiments, this invention can be applied to a computer-implemented system for rendering a stream of images, such as the Windows Media Player, but can be applied to personal video recorders, such as the Sky + (TM) box. , TiVo recorder, portable video player, such as an iPod video, or DVD players or video players. In many such systems, the image stream may originate from a broadcast, a cable television transmission, a Web direct playback Internet server, or the like, and is recorded prior to playback using a suitable player. If the video stream is directed one by one from an Internet server to a user, then this search tool could operate at the end of the Web server, but it is controlled by the user. Thus, if the user wants to search for a particular point within the video stream, or wants to jump to a particular part of the video stream, he can press a button or key that causes the system to go into a search mode whereby the user can search for a start image according to the invention. Thus, in that embodiment, the Web server will generate a set of candidate images to be rendered on a user's machine, which could be a PC or Internet-enabled television or the like. The user can then select the closest candidate image, and the remote Web server will then render another set of candidate images for further selection by the user. Clearly, this not only applies to direct video playback over the Internet, it can also be applied to other systems where a remote server serves video content on a one-to-one basis rather than a broadcast to multiple users.

In a second aspect of the invention, there is a system as defined in claim 23.

Advantageously, the system allows a viewer to quickly identify the point at which he chooses to start or restart watching a program. It is preferable that the processor is further arranged, in use, to reduce a range of the image stream sought in each of the successive iterations. It is further preferable that, at each iteration, the controller is set to use to re-select one of the candidate images displayed on the screen as the reference image. Furthermore, the screen may be arranged to represent the candidate images chronologically and the reference image may be arranged to precede one or more of the candidate images in the image stream.

It is preferable that the system is arranged to begin reproducing the image stream from a candidate image in response to a firm selection of that candidate image. For example, the system may be arranged so that firm selection is achieved by holding down a key on the controller for a relatively long duration. This allows a user to save time and skip additional iterations if they are satisfied for the image stream to be displayed from the candidate image that is firmly selected. Additionally, the processor may be arranged to allow adjusting the time interval between candidate images. In one arrangement, the controller includes one or more keys arranged to be operated by a user to adjust the time interval between candidate images, for example, the increase keys.

ES 2 395 085 T3 volume and decrease the volume. In the preferred embodiment, the processor is arranged to cause the display to display the time interval for each of the candidate images from the initial reference.

The controller may include at least two keys configured in the same arrangement that the candidate images are displayed on the screen, each candidate image corresponding to a key. The processor may be arranged to select the candidate images, each candidate image being spaced at the same time interval in the image stream, and reducing the time interval for each successive iteration and in the preferred embodiment, the reference image is re -select a predetermined number of times. The processor may be arranged to control the display means to display the image stream starting from the start image. The processor may be arranged to analyze the time interval between candidate images or the duration of the image stream, and increase or decrease the number of iterations executed by the procedure accordingly. This allows the system to be able to choose an appropriate number of iterations depending on the length of the image stream it is looking for. If a user is simply skipping 3 or 4 minutes of ads, then fewer interactions are likely to be needed than if a user is searching through a 5 hour video recording.

The system may also further comprise a cache memory, the cache memory being controllable by the processor and being arranged to download a part of this image stream locally to the processor. Advantageously, the flow of images that is required to be displayed on the screen, either as a continuous image or the candidate image, is readily available on the cache. This avoids the delay in the request, reception and obtaining of the part of the flow of images for its immediate representation.

The system can be a personal video recorder that includes the processor and a memory. Preferably the memory is hard disk or solid state memory.

Various additional aspects of the preferred features will be apparent from the appended claims.

Additional features and advantages of the present invention will be apparent from the following description of the embodiments thereof, presented by way of example only, and with reference to the accompanying drawings, in which:

Figure 1 is a schematic diagram depicting essential components of a preferred embodiment of a system for the invention;

Figure 2 is a schematic diagram depicting a keyboard of a controller for use in the invention;

Figure 3 is a schematic flow diagram depicting the steps in a process incorporating the invention;

Figure 4 is a schematic representation of a representation that has nine candidate images in a first iteration:

Figure 5 is a schematic representation of a representation having nine candidate images in a second iteration;

Figure 6 is a schematic representation of a representation having nine candidate images in a third iteration;

Figure 7 is a schematic representation of a representation showing nine candidate images in a first iteration, and having the image stream shifted back by one step;

Figure 8 is a schematic representation of a representation having nine candidate images in a first iteration, and having the image stream moved forward one step; and Figure 9 is a diagram showing a computer implementing a further embodiment of the invention.

Figure 1 illustrates an example of a suitable system environment including a Sky + ™ Box, in which the invention can be implemented. This system environment is only one example of a suitable system environment, which is the preferred embodiment, and is not intended to suggest any limitation to the scope or functionality of the invention.

Example system 1 for the implementation of the invention shown in Figure 1 has components including: a data storage 3 for supplying an image stream; a processor 5, a cache memory 7, a display unit 9; and a controller 11. Processor 5 and cache 7 are connected to each other and are each arranged to communicate with data storage 3. The cache is connected to the display unit 9 and the controller is arranged to communicate wirelessly, or by a wired means, with the processor 5.

In use, a request is sent from processor 5 to data store 3. Data store 3 downloads a stream of images to cache memory 7. Once instructions are received from processor 5, the

ES 2 395 085 T3 cache memory 7 redirects the image stream to the display unit 9 for rendering the image stream. Consequently the cache memory 7 can download and store the image stream in advance of the presentation of the image stream on the display unit 9.

Since the controller 11 can be operated to control the processor 5, the controller 11 can be used to control the flow of images that is displayed on the display unit 9, for example to pause the flow of images on a particular image, or for fast forwarding or rewinding of the image stream representation.

Figure 2 shows a controller 11 suitable for use in system 1. Controller 11 has a numeric keypad for the numbers 1 to 9, (represented by label 30) and for zero (32). It also has a channel up button 34, a channel down button 36, a back button 38, and a volume control button 40.

Upon receiving the instructions from the controller 11, the processor 5 is arranged to control the data storage 3 to supply the image stream to the cache memory 7 to store the data stream and to redirect parts of the stored data stream to the unit. display 9. The processor is also arranged to receive the signals from the cache memory 7. These signals include acknowledgments from cache 7 that indicate whether it is storing a particular part of the image stream.

Embodiments of the invention can be described in the general context of computer executable instructions, such as program modules that are executed by a computer. In general, program modules include routines, programs, objects, components, data structures, and so on, that perform particular tasks. The invention can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communication network. In a distributed computing environment, program modules can be located on a local and / or remote computer storage medium, including memory storage devices. In this particular system, the data storage may be a device remote from the rest of the system (for example at the head end of a cable television system), and the task of supplying the image stream is performed remotely.

In addition to the general functions described above, the system incorporating the invention is arranged to implement a procedure for selecting the starting point in the image stream at which the representation of the image stream will begin, as shown schematically in the diagram of flow of Figure 3. The method enables a viewer to iteratively search for the start image in an image stream of a recording delivered from a computer-readable medium, without having to fast-forward or rewind the representation of the image stream.

In the flowchart, the solid single headed arrows indicate the movement of the procedure from one stage to the next where there are no alternatives to that movement. An open single headed arrow indicates that of the two alternative results for a stage, the arrow is following a negative result; and an open double headed arrow indicates that the result of the stage is a positive result.

To find a starting point in the image stream, you first need to select or define the beginning and end of the image stream that you are looking for. In some cases, this will correspond to a data packet or the entire recording. For example, on a DVD that contains multiple episodes of Friends, the data stream to search for might be one of the episodes. Alternatively, it could be part of an episode depending on the length of the data packets. If the DVD is a movie, such as Star Wars, the movie can be divided into multiple packages, each of which could be searched independently. Alternatively, you could take the entire movie and search through the entire movie. In a case where the recording has been made from a television broadcast, and it is desired to skip the commercials that have been recorded, the video stream to search could be defined as a period of five minutes following the point in the program that was being watched when the jump was activated.

In activating the search facility, and selecting the image stream to search for, nine images are presented to the viewer in a three-by-three arrangement on the television screen. Of course, in other embodiments, the image arrangement could be larger or smaller, but nine is a convenient number that corresponds to the nine natural numbers on the keyboard.

Generally speaking, each of the nine images displayed in the image array are spaced within the image stream. In this way, the first image will be at or near the beginning of the image stream, and each subsequent image will be evenly spaced through the rest of the image stream. In the described embodiments, the image stream is effectively divided into eight parts, and with the exception of the ninth image, each of the images within the arrangement generally corresponds to the beginning of each of the eight parts. Each of the images represented in the layout is a candidate image that can be chosen by the user by pressing the corresponding key on the keyboard. In the embodiment shown in Figure 4, the nine images are spread out over a period of only five minutes. Thus, it can be used to skip viewing a commercial intermission within a movie or television show that has been recorded. To skip the intermission, the user simply selects the candidate image which is the first image after the ad intermission and which immediately brings the viewer to a position fairly close to the end of the ad intermission. A new arrangement of nine candidate images will be

ES 2 395 085 T3 then represent on the screen corresponding to the part of the flow of images that was selected by the user in his first selection. Thus, the new set of candidate images forms a second iteration of the procedure where the first image not associated with the ad intermission is again selected, thereby bringing the viewer closer to the end of the ad intermission. By performing this action in several iterations, the user arrives very quickly on the starting image from which the viewing of the program or movie should begin again. The number of iterations is likely to depend on the initial length of the image stream. The number of iterations is probably less to skip an ad intermission than to find a point within a two-hour movie. The processor may include an algorithm that determines the number of iterations based on the length of the image stream being searched through, or the length of the time interval in the current iteration.

Selection of the appropriate candidate image is accomplished above by pressing the corresponding key on the keyboard. Selection is achieved by a short normal press of the key, However, if the user wishes to simply start the image flow from one of the images without going to an additional iteration, they can press the appropriate key corresponding to the image from which The flow of images should continue and you can hold down the key for a longer period of time, for example 1 second. This kind of pick can be described as hard pick that refers to how long the key is held down, rather than how hard it is pressed. A firm selection could also be achieved by some other means than by selecting the image. For example, the numbered key on the keyboard could be pressed in conjunction with a shift key or an extension key that might be present (not shown). Instead of selecting the candidate images using the numbered keys on a keyboard, an on-screen selection system can be used, in which the candidate image is selected on the screen. Typical examples of a selection on the screen would be a movable pointer that can be positioned over the candidate image to be selected, and then selected by the click or enter button on the keyboard. Another type of on-screen system highlights each of the images by the use of a focus, and the arrow keys or a control knob on the keyboard can be used to move the focus from candidate image to candidate image. The candidate image can be selected, for example, by pressing an enter key. In such cases, a hard selection could be made by pressing an alternate key, or by a double click on the selection or the enter key instead of a single click used to make a soft selection. In a further embodiment (not shown), the screen is touch sensitive, and to make a selection the user simply touches the appropriate candidate image on the screen.

More details of embodiments of the invention will now be described.

Initially, the image stream is displayed on the display unit 9. A viewer initiates, in a first step 101, a procedure implementing the procedure by pressing a search key on the controller 11. This selects an initial reference image. It could be the image displayed on the screen at the moment of pressing the search key, or the image of a short time before the pressing of the key. In the example below, the image stream is imaged five seconds before the key is pressed. The search that starts at this point is the one that allows the user to skip the ads. Therefore the user starts the search when he reaches the beginning of the ad interval and allows him to skip it quickly and easily. The initial reference time generally corresponds to the beginning of the announcement interval. The search will take place over the next 320 seconds of the image stream. In the method all additional images in the image stream have a time interval T with respect to an initialization image. The initialization image is the image that was rendered when the search key was pressed to start the search facility. The reference image therefore has a time interval T of -5 seconds. Having a time interval of -5 seconds, the entire video stream interval is slightly extended. This allows the first candidate image to be fixed slightly earlier than the time at which the search was selected. In some embodiments, users might find it more acceptable if the time is set to 0 seconds. This implementation will be described with reference to Figure 4.

In a second stage 103, the processor sets the conditions for the search. Each of the iterations has a count value (C), which is initially set to one. At each iteration, the processor selects a number of candidate images. The reference time of the first iteration Tr1 (Trc, where C is the value of the iteration count) is set to the same value as the first candidate image 13, T1.

In a third stage, 105, the processor 5 instructs the cache memory 7 to provide nine candidate images 15 of the image stream. Each of the images is spaced in the image stream of the adjacent images by a time interval, ot, of 40 seconds. This value could be changed by the user if required. For example, the value could be changed 131, 132, using the volume keys 40. Instead of increasing and decreasing the sound, the volume keys could be used to increase 131 and decrease 132 the time interval 133, 134. Alternatively, the value of the time interval 133, 134 could be changed using some other input from the keyboard or elsewhere. In Figure 3 in boxes 133, 134, when the volume down or volume up keys are pressed, the time interval is either doubled or halved in a geometric progression. However, alternative embodiments could increase or decrease the time interval by a fixed amount of time, for example by 20% of the original time interval. This will increase or decrease the time interval in stages that are in an arithmetic progression rather than a geometric progression. ót is the time space between images, that is, between T1 and T2, T2 and T3, etc. This is dependent on the

ES 2 395 085 T3 C value. In this embodiment;

to

<td>C = 1</td><td>ot = 40 s</td>

<td>C = 2</td><td>ot = 5 s</td>

<td>C = 3</td><td>ot = 0.625 s</td>

The third of these nine images is the first candidate image 13. If the cache does not store any of these images, the processor 5 instructs the data storage 3 to send the image stream that has these lost images to the cache 7. The cache memory 7 then sends the candidate images 15 to the display unit 9. The images are displayed on the display unit in chronological order. The first image in T₁, in the image stream is displayed as the first candidate image 13, in the upper left corner of the display 17 of the display unit 9. The second image at T2 in the image stream is displayed in the middle of the row top on the screen 17. This is repeated for each of the remaining seven images so that the last T9 candidate image in the image stream is the image in the lower right corner of screen 17, as shown in Figure 4. Candidate images are displayed. they represent the viewer until the processor 5 receives a signal from the controller 11. The processor 5 waits for a signal from the controller 11 in the fourth step 107. Any one of several different inputs can be received from the controller 11. This is clear from Figure 3. However, it is ultimately necessary for the user to select one of the candidate images which is designated in this case as the fifth step 109. Without However, it will be appreciated that between the fourth and fifth stages, there are other inputs that can be made.

The controller has a keyboard that is a standard numeric keypad. Each of the number keys 30 is between one and nine, inclusive, and has an associated candidate image 13. A viewer selects one of the new candidate images. The reference image for the next iteration will be the candidate image before the one selected (pressing the 6 key selects the fifth image as the reference image for the next iteration). In a sixth step 109, where the key is not a number key 30 between 1 and 9, the procedure returns to step 107 until a number key is pressed. As described above, if a firm selection is made, the video stream will play from the selected image. In this embodiment, a firm selection is made by a double click or a double click on the number key, and this is identified by the fifth step 135. If a double click is not detected, then the procedure moves to step 6. However, if a double input is detected, then playback 136 resumes.

In the sixth step 109, the processor writes down the input number key, α. In a seventh step 111, the iteration count C is checked to ensure that it is not equal to three, indicative of a third iteration. (The processor is ready to stop after three iterations.) If C is not equal to three, the procedure continues to a seventh step 113 to continue the procedure in a further iteration in which the time interval ot is reduced.

In an eighth step 113, the search conditions are altered. The value of account C is increased by one. The instant in the image stream of the new first candidate image T is calculated₁, that is, the new reference image, whereby the current time interval, ót, is multiplied by an integer that is one less than the value of the pressed number key, α, and this value is then added to the time from the first candidate image above:

^T1 new = ^T1 previous ^{+ ((α - 1) x ót)}

The value of the reference time for the second iteration Tr2, is set to the new value of the first new candidate image T1. The procedure then returns to the third step 105. The time interval, ot, is reduced by a factor of eight, although this could be any value up to one less than the number of candidate images selected. The display unit 9 represents nine candidate images as shown in Figure 5, where the nine images are each spaced along the image stream by a time interval of five seconds. The selection of the candidate images shown in Figure 5 follows from the selection of the seventh image of the first iteration shown in Figure 4. In Figure 5 the first candidate image 19 is located in the image stream 195 seconds after the initialization image.

The third, fourth, fifth, sixth, seventh and eighth steps of the procedure are repeated, as shown by way of example in Figure 6 for the third iteration. In the second iteration of this example, the number key six is ​​pressed, and in the third iteration, the number key seven is pressed. In the third iteration the candidate images are spaced by 0.625 second intervals. Where, at the seventh step 111, the iteration count C is found to be three, the procedure then proceeds to the ninth step 115. The candidate image before the selected image becomes the start image.

In the ninth step 115, the start image is displayed on the screen. In one embodiment the display unit 9 starts the reproduction of the image stream immediately. In another embodiment, the image stream is frozen on display unit 9 until a key on controller 11 is pressed to start playback.

ES 2 395 085 T3

In this preferred embodiment, playback of an image stream can be resumed without a user selection of the candidate images. If the backspace button 38, which could also be the zero number key 32, is pressed on the controller number pad 11, the procedure enters a backspace mode 117. If the iteration count C has a value equal to one, in step 119, the image stream is reproduced from one second in the image stream before the initialization image, as in step 121. If the iteration count , C, is not equal to one, in step 123, the reference image is the same as the first candidate image of the previous iteration, and the value of the iteration count C is reduced by one. The procedure then returns to the third step 105, in which the display unit 9 represents the candidate images from the previous iteration. This also enables a viewer to review the candidate images from the previous iteration if he realizes that he has selected the wrong candidate image in one iteration.

Alternatively or additionally, if the viewer selects candidate images that have an interval that does not include the start image, they can search through the image stream without going back to the previous iteration. If the start image is before the first candidate image on the display unit 9, at T9 in the image stream, the viewer could press a channel upload button 34 on the controller 11. The procedure enters a scrolling up mode 125. At this step 127 the time of the first candidate image T1 is reset to precede the current value of T1 by six time intervals, ot, of the current iteration. The time value of the current iteration Trc is set equal to the new value of T1. The procedure returns to the third step 105, the display unit 9 represents the nine candidate images, each spaced in the time interval ot, and where the first candidate image is taken from the image stream in the new T₁. Figure 7 shows the effect of the scrolling procedure of the image shown in Figure 4.

If the viewer chooses to search for the starting image after the last currently rendered candidate image, the viewer could press a channel down button 36 on the controller keyboard 11, to enter a scroll down procedure 129. At this step 130 , the time of the seventh candidate image at T7 becomes the time of the first candidate image at T1. The time value of the current iteration Trc takes on the new value of T1. The procedure then returns to the third step 105, as shown in Figure 8, where the procedure of scrolling down has been applied to the images represented in Figure 4.

In the use of the invention when playing back a recording, the user can select an initial reference image 101. The screen displays nine images 40 seconds apart and the first image T1 precedes the initialization image by five seconds, as shown in Figure 4. If the viewer were to select the seventh image 21 in Figure 4, the display unit 9 would then render nine images 5 seconds apart, as shown in Figure 5, the last image being the seventh image 21 in Figure 4 . If the viewer then selects the sixth image 23, the display unit 9 represents the nine images as shown in Figure 6, where the last image T9 is the sixth image 23 of Figure 5, and the images are 0.625 apart. seconds. The viewer can select their start image from any of these nine images. The starting image will be the candidate image that precedes the image that is finally selected. The display unit 9 will then start rendering the image stream from the start image.

In all these candidate images, the position of each image in the image stream is expressed as a function of time with respect to the initialization image, and can be represented on the display unit 9 in association with the respective candidate image.

This preferred embodiment searches for the start image by selecting an image to be the first image in the range of the image stream that the viewer chooses to view. The procedure can be modified so that the image selected for each of the iterations becomes the first image of the next iteration, so that the procedure selects the starting image by means of the last image in the interval of the stream. images that the viewer chooses not to see.

The numeric key values ​​associated with the image can also be represented with the associated image. Any number of images can be displayed on the display unit 9, as long as at least two are displayed, and the viewer has a means to select all the images in the image stream. The initial time interval between candidate images can be any value appropriate for the purpose of the search, and the number of iterations can also be any predetermined number. In effect, the number of iterations need not be predetermined and may be conditional on the size of the time interval between candidate images in an iteration.

The proportion of the time interval between iterations can be any value, as long as there are a sufficient number of candidate images, and the time interval between them is sufficient, for the entire chosen interval of the image stream to be searched.

Information about each of the candidate images, for example the corresponding key on the keyboard and the location of time in the image stream with respect to the initial reference image, can be represented so that this information does not obscure the images. .

Images can be still or moving. If the images are moving the duration of each clip of

ES 2 395 085 T3 image is 2 seconds, but could be more or less than this in other embodiments. Once the time interval is shortened so that it is shorter than the time interval between candidate images in the image stream, each of the images is rendered for the duration of the time interval, with each of the images starting from one of the candidate images. To ensure that the image of the chosen starting point is seen, the viewer will select a candidate image as their reference image that precedes the starting image that they would normally choose if the images were still.

The image stream may comprise a series of frames. If the time in a selected image stream from an image does not have an image frame, or has a contrast that is insufficient for the image to be discernible, the next frame in the image stream is rendered.

The controller can be operated by keys or means of operation other than the numeric keys. The arrangement of the candidate images may be arranged to conform to the arrangement of the keys used in the controller.

Additional search modes may be used within the scope of the invention. In a further embodiment that can be run on top of the system shown in Figure 1, the processor can be arranged so that when a user searches forward or backward through a stream of images while viewing the stream of images (signal / review), when the user stops searching, a multi-image search mode is entered. In conventional machines for viewing video programs or recordings, it is possible to search forward or backward while the stream of images is being viewed, and it is often possible to search at different speeds. Such back and forth searching is normally done to find a particular point at which the user wishes to begin viewing a recording. In such a search, an amount of excess is unavoidable: it takes a certain amount of time to stop the search once the point from which you want to start is recognized. The excess will depend on two things, firstly the user's reaction speed, and secondly the speed at which the video stream is being searched. The faster the search, the greater the excess. In this search mode, when the search is stopped, the processor in the device is arranged so that it enters the multi-image search mode in which several images are put on the screen, preferably nine, which are displayed on the screen. order in which they will appear in the image stream. When the multi-image search mode that follows a backward search through the image stream has been entered, the final image that is obtained at the end of the excess forms the first candidate image. Eight after images are displayed, each of which is separated from the previous image by a time interval or a space of time. The time interval could be dependent on the speed with which the forward search or the reverse search was performed immediately before the search was stopped. Therefore, a fast search might require a 10 second time interval between each of the images, while a slower search will result in a smaller overshoot, and the time interval could be 1 second. However, these are examples and the time interval could be longer than 10 seconds or less than 1 second.

If the search before the excess was a forward search, then the image selected at the end of the excess will form the last of the candidate images, and the other eight images will be arranged sequentially before the final candidate image. Selection by the user of the most appropriate candidate image may then allow the image stream to be reproduced from that point, or alternatively result in one or more additional iterations.

Some personal video recorders that are available can be arranged to record the last 10, 15, or an alternate number of minutes of a live broadcast or other stream of images that a viewer is viewing. In a further search mode, which could be called a replay action search mode, the processor is arranged so that when the viewer is watching a broadcast or other image stream, the period of the viewed image stream is recorded. recently, for example the previous 10 minutes of the image stream, and the processor is further arranged so that when the search button is pressed, collects nine candidate images from within the recording time. The ninth image will be the one that was currently in when the search button was pressed. If the spacing is 10 seconds, the maximum length of the image stream that should have been recorded will be 80 seconds. Pressing the number on the keyboard relative to the candidate image that was selected will cause the video stream to start from the selected candidate image, effectively giving an ease of repeatability of the action. It may be possible to increase or decrease the time interval between candidate images using the volume control buttons. If it is possible to increase the time interval, then more than 80 seconds will need to be recorded from the previously viewed image stream. Additional iterations should also be possible, if required. The system 1 on which the invention operates is whatever is suitable for reproducing a recording as a stream of images, for example a CD Rom player, a DVD player, a computer running a video player program or a Sky + ™ Box. Examples of computer video players are Microsoft's Media Player and Adobe's Flash Player and RealPlayer. The software that provides the functionality of the present invention may be incorporated into the computer video player software, or it may be provided by means of add-on software that adds the functionality to an existing computer video player program that does not originally has the functionality of the present invention. The data store 3 can be connected to the processor 5, or it can be remote from the rest of the system 1. The arrangement of the programmer 5, the cache memory 7, the data storage 3 and the display unit 9 described so far are exemplary only. Processor 5 could communicate exclusively with data storage

ES 2 395 085 T3

3. The cache memory 7 could be remote from the data storage 3 and the processor 5, if placed with the display unit 9. The display unit 9, the processor 5 and the data storage 3 may therefore each comprise a computer that includes a computer-readable medium.

The computer-readable medium can be any available medium that can be accessed by the computer and includes both volatile and non-volatile media, and removable and non-removable media. By way of example, and not limitation, the computer-readable medium may comprise a computer storage medium and a communication medium.

Computer storage medium includes volatile and non-volatile media, removable and non-removable media implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-Rom, digital versatile discs (DVD) or other optical disk storages, magnetic disk storages or devices. magnetic storage or any other medium that can be used to store the desired information and that can be accessible by the computer.

Communications media typically incorporate computer-readable instructions, data structures, program modules, or other data into a modulated data signal such as a carrier wave or other transport mechanism, and include any information delivery medium. The term "modulated data signal" means a signal that has one or more of its characteristics fixed or changed in such a way as to encode the information in the signal. The term "image stream" is referred to herein as a modulated data signal that encodes the image data. By way of example and not limitation, communication media includes wired media such as a wired network, or a direct wired connection, and wireless such as acoustic RF, infrared, and other wireless media. Combinations of any of the above should be included within the scope of computer-readable media.

The image stream may also have an audio component and the system is arranged to broadcast that audio component together with the representation of the image stream on a display unit.

A computer-implemented embodiment is shown in Figure 9, in which the computer system 60 is shown including a PC, a computer screen 62, a keyboard 63, and a controller 69. The PC includes a media control unit, which in this example is a DVD 64 player. It will be appreciated that any suitable driver could be included, or the hard disk of a PC could be used as the source of the image stream. Additionally, the image stream could be received from any other suitable source, including the Internet 70. Running on the PC 61 is an operating system 65 on which a video player 66 program is running, such as the Video Media Player. Microsoft, an add-on software to the search tool 67, and performance data on which the video player program 66 operates. In this example, the operating data would normally be the video stream. The search tool companion software 67 is a piece of software in addition to the video player program, although it could be incorporated within the video player program itself. Thus, the add-on software can be supplied as standalone software, or it can be incorporated into the video player program as required. The combination of the video player program and the search tool add-on software operate together means that, during the use of the video player program to render a video stream, the add-on software 67 can be invoked to use the search facilities described in Figures 3 to 8. A controller 69 could be used, or the functions of the controller buttons could be replaced by keys on the keyboard 63, buttons available on a touch screen, or buttons that can be selected using a mouse (not shown).