JP2009177822A - Specific image detector, specific image detection method, program, and specific image detection system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently detect an objective specific image from continuous images. <P>SOLUTION: In this specific image detector 20, an input frame image and a frame image delayed through a delay circuit 5 are supplied to a quiescence determination part 6, and the input frame image is detected as a candidate image when a ratio of luminance changes of both the images is below a predetermined threshold. Image feature value acquisition parts 3 and 7 acquire feature quantities of the input frame image (registered image) and the candidate image when a user instructs registration. The feature value of the registered image is stored in a feature value table 4. A similarity determination part 8 determines the degree of similarity between the feature value stored in the feature value table 4 and the feature value of the candidate image, and displays it on a sub-monitor 11 through a presentation control part 9 as a specific image when they are similar to each other. The specific image detector 20 has a plurality of tuners, displays an image of a reception channel selected by one-side tuner on a main monitor 12, and detects a specific image from images of a reception channel selected by another tuner to display it on the sub monitor 11. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a specific image detection apparatus, a specific image detection method, a program, and a specific image detection system for detecting a specific image from a still image included in a video composed of continuous frame-unit images.

  2. Description of the Related Art Conventionally, there is a technique in which information that is important for a user (user) is registered in advance and an image is searched from the characteristics of the registered information.

  For example, Japanese Patent Laid-Open No. 2000-308017 (Patent Document 1) discloses a video viewing apparatus that efficiently views a related video in a limited time. The video viewing device described in this publication reproduces a predetermined input video signal stored in the input video storage means from a predetermined position, and an input video storage means for storing an input video signal including a moving image signal and an audio signal. Video reproduction means, calculation means for calculating the feature value of the input video signal, feature value storage means for storing the feature value calculated by the calculation means, and feature values from the calculation means are features of the feature value storage means A video search unit that searches for video related information indicating that these feature values are related, and a related information storage unit that stores video related information; When the related information storage means stores the related information, it notifies the viewer of the presence of the related information and notifies the viewer, and requests the reproduction of the video signal corresponding to the feature value indicated by the notified related information. Input means for inputting a request signal, and a insertion means for reproducing the requested video signal. Thereby, it is possible to detect a video signal related to the currently viewed video from the stored video signals and view the related video signal, and efficiently view the related video in a limited time Can do.

  Japanese Patent Laid-Open No. 10-149373 (Patent Document 2) discloses an image storage / management apparatus and an image for searching an image database system that stores and manages a large amount of image information in a computer and performs search / use. An indexing method is disclosed. In the image storage / management apparatus described in this publication, input means for inputting an original image, storage means for storing the input original image, and area division information by dividing an area based on visual characteristics of the original image Area dividing means for generating the original image, block dividing means for dividing the original image into a grid and generating division information, and inputting the division information, determining and arraying the representative area number for each grid, and dividing the area Index generation means for extracting area attribute information including color information for each area and describing image index information; and storage means for storing image index information. Thereby, it is possible to search for an image using a visual clue that the user has with respect to the image.

JP 2000-308017 A Japanese Patent Laid-Open No. 10-149373

  However, in conventional techniques such as the technique described in the above-mentioned publication, a method for efficiently detecting a target image from changes in images of various patterns has not been established. In other words, in various patterns of images, the physical quantity indicating the characteristics that identify the target image is not obtained, and the target image cannot be accurately searched. There is a problem that it cannot be detected efficiently.

  The present invention has been proposed in view of such a conventional situation, and a specific image detection apparatus and a specific image detection capable of detecting a target specific image with high efficiency from a continuous image in real time. It is an object to provide a method, a program, and a specific image detection system.

  In order to achieve the above-described object, a specific image detection apparatus according to the present invention is a specific image detection apparatus that detects a specific image including a still image part included in a video composed of images in consecutive frames. The ratio of the luminance change between the storage means for storing one or more feature quantities having different characteristics in the registered image and the frame image and the adjacent frame image close in time to the frame image is calculated, and the luminance change ratio is predetermined. A candidate image detecting means for detecting a frame image that is equal to or less than a threshold value as a candidate image, a feature quantity detecting means for detecting one or more feature quantities having different characteristics in the candidate image, a feature quantity of the candidate image, and the registration When the similarity with the feature amount of the image is equal to or greater than a predetermined threshold, a detection unit that detects the candidate image as a specific image, and the specific image detected by the detection unit are displayed on the display unit. Presenting means and at least two tuner means, wherein the presenting means displays the video of the reception channel selected by one tuner means on the display means, and the detecting means is the other tuner means. The specific image is detected from the video of the reception channel selected by.

  The specific image detection apparatus according to the present invention is a specific image detection apparatus that detects a specific image including a still image portion included in a video composed of images in consecutive frames, and has different characteristics in a registered image. The frame means for calculating the luminance change ratio between the storage means for storing the above feature amount and the frame image and the adjacent frame image temporally adjacent thereto, and the luminance change ratio is equal to or less than a predetermined threshold value A candidate image detecting means for detecting a candidate image, a feature quantity detecting means for detecting one or more feature quantities having different characteristics in the candidate image, a feature quantity of the candidate image, and a feature quantity of the registered image A detection means for detecting the candidate image as a specific image when the degree is equal to or greater than a predetermined threshold; a presentation means for displaying the specific image detected by the detection means on the display means; Also includes two tuner means, and the detection means detects the specific image from the video of the reception channel selected by one of the tuner means, and the presentation means is a first display means and a second display means. The video of the reception channel selected by the other tuner means is displayed on the first display means, and the specific image is displayed on the second display means.

  Furthermore, the specific image detection apparatus according to the present invention is a specific image detection apparatus that detects a specific image including a still image portion included in a video composed of images in consecutive frames, and has different characteristics in a registered image. The frame means for calculating the luminance change ratio between the storage means for storing the above feature amount and the frame image and the adjacent frame image temporally adjacent thereto, and the luminance change ratio is equal to or less than a predetermined threshold value A candidate image detecting means for detecting a candidate image, a feature quantity detecting means for detecting one or more feature quantities having different characteristics in the candidate image, a feature quantity of the candidate image, and a feature quantity of the registered image A detecting unit that detects the candidate image as a specific image when the degree is equal to or greater than a predetermined threshold; and a presentation unit that displays the specific image detected by the detecting unit on a display unit. It said presenting means, the specific image detected by said detecting means a plurality of holding, at the same time displayed on the display means by reducing the specific image of the plurality of.

  In the present invention, a frame image having a luminance change rate equal to or less than a predetermined threshold is detected as a candidate image, and a candidate image similar to a registered image registered in advance by a user or the like is detected as a specific image. The registered image can be selected and detected with high efficiency because of its importance, and the frequency of false detection can be reduced because the image is specified by the feature amount after the candidate image is detected. .

  In addition, the feature amount stored in the storage unit includes luminance information, color information, edge amount, edge density, luminance histogram, stationary time of the registered image, and video of a predetermined period including the registered image. One or more pieces of information regarding the included sound effects can be included, and the registered image can be detected more accurately by determining the degree of similarity using feature quantities having a plurality of different characteristics. For example, when a specific image is detected from a television image, the robustness is improved even when the specific image is detected from a noisy television image by using a luminance histogram and a still time for determining the similarity. can do.

  Further, the registered image is an image designated by a user, and the feature amount detecting means has one or more feature amounts having different characteristics in the registered image, which is an image designated by the user, and the storage means The feature amount of the registered image detected by the feature amount detecting means can be stored, and the user can register the registered image as needed.

  Furthermore, the storage means has data consisting of the feature amount of the registered image and the similarity threshold, and an input means for inputting a user's evaluation for the specific image detected by the detection means; Rewriting means for rewriting the similarity threshold of the storage means based on the user's evaluation input via the input means, and the similarity threshold is rewritten by the user's evaluation. False detection is reduced.

  Moreover, it can have a presentation means which displays the specific image detected by the said detection means on a display means.

  Further, connected to an external device that supplies video, the detection means can detect the specific image from the video supplied from the external device.

  Furthermore, the apparatus has at least two tuner means, the presenting means displays the video of the reception channel selected by one tuner means on the display means, and the detecting means is selected by the other tuner means. The specific image can be detected from the received video of the receiving channel, and the user can view the television video while detecting the specific image.

  Also, it has at least two tuner means, and the detection means detects the specific image from the video of the reception channel selected by one of the tuner means, and the presentation means includes a first display means and a second display means An image of the reception channel connected to the display means and selected by the other tuner means can be displayed on the first display means, and the specific image can be displayed on the second display means. While viewing the television image on one display means, a specific screen is detected from a television image or the like of a channel different from the television image that the user is viewing, and the detected specific image is detected by the other display means. Can see.

  Furthermore, the detection means can detect the specific image from videos of a plurality of channels selected in time division.

  Furthermore, the presenting means may hold a plurality of the specific images detected by the detecting means, reduce the plurality of specific images, and display them simultaneously on the display means.

  Further, the image processing apparatus includes block dividing means for dividing the frame image into blocks of a predetermined size, and the detection means has a luminance change between corresponding blocks in the frame image and the adjacent frame image being equal to or less than a predetermined threshold value. At this time, the frame image can be detected as a candidate image.

  Furthermore, it has a luminance histogram calculating means for calculating a luminance histogram for each block divided by the block dividing means, and the candidate image detecting means is a luminance histogram between blocks corresponding to the frame image and the adjacent frame image. When the number of blocks whose correlation coefficient is equal to or smaller than a predetermined threshold is equal to or smaller than the predetermined threshold, the frame image can be detected as a candidate image.

  Furthermore, a notification means for notifying the user when a specific image is detected by the detection means may be provided, thereby improving the convenience for the user.

  A specific image detection method according to the present invention is a specific image detection method of a specific image detection apparatus for detecting a specific image including a still image portion included in a video composed of images in units of continuous frames. A candidate image detection step of calculating a ratio of a luminance change between the image and a close frame image that is temporally close to the image, and detecting a frame image in which the ratio of the luminance change is equal to or less than a predetermined threshold; A candidate image feature amount detecting step for detecting one or more feature amounts having different characteristics in the candidate image, a feature amount of the candidate image, and one or more feature amounts having different characteristics in the registered image stored in the storage unit A detection step of detecting the candidate image as a specific image when the similarity is equal to or greater than a predetermined threshold; and a presentation process for displaying the specific image detected in the detection step on a display means The specific image detecting apparatus has at least two tuner means, and in the presenting step, the video of the reception channel selected by one tuner means is displayed on the display means, and the detecting step Then, the specific image is detected from the video of the reception channel selected by the other tuner means.

  A specific image detection method according to the present invention is a specific image detection method of a specific image detection apparatus for detecting a specific image including a still image portion included in a video composed of images of successive frames. A candidate image detection step of calculating a luminance change ratio between a frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image A candidate image feature amount detecting step for detecting one or more feature amounts having different characteristics in the candidate image, a feature amount of the candidate image, and one or more feature amounts having different characteristics in the registered image stored in the storage unit When the similarity is equal to or greater than a predetermined threshold value, a detection step of detecting the candidate image as a specific image and a specific image detected in the detection step are displayed on the display means The specific image detecting apparatus has at least two tuner means, and in the detecting step, the specific image is detected from the video of the reception channel selected by one tuner means, In the presenting step, the video of the reception channel selected by the other tuner means is displayed on the first display means, and the specific image is displayed on the second display means.

  Furthermore, the specific image detection method according to the present invention is a specific image detection method of a specific image detection apparatus for detecting a specific image including a still image portion included in a video composed of images of successive frames, A candidate image detection step of calculating a luminance change ratio between a frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image A candidate image feature amount detecting step for detecting one or more feature amounts having different characteristics in the candidate image, a feature amount of the candidate image, and one or more feature amounts having different characteristics in the registered image stored in the storage unit When the similarity is greater than or equal to a predetermined threshold value, a detection step of detecting the candidate image as a specific image and a specific image detected in the detection step are displayed on the display means To execute the presentation step, in the present process, the specific image detected by said detecting step to keep multiple simultaneously displayed on the display means by reducing the specific image of the plurality of.

  A program according to the present invention is a program for causing a computer to execute an operation of a specific image detection device that detects a specific image from a still image included in a video composed of continuous frame-unit images. A candidate image detection step of calculating a ratio of a luminance change between the image and a close frame image that is temporally close to the image, and detecting a frame image in which the ratio of the luminance change is equal to or less than a predetermined threshold; A candidate image feature amount detecting step for detecting one or more feature amounts having different characteristics in the candidate image, a feature amount of the candidate image, and one or more feature amounts having different characteristics in the registered image stored in the storage unit When the similarity is equal to or greater than a predetermined threshold, a detection process for detecting the candidate image as a specific image and a specific image detected in the detection process are displayed. The specific image detection apparatus has at least two tuner means, and in the presentation step, the video of the reception channel selected by one tuner means is displayed on the display means. In the detection step, the specific image is detected from the video of the reception channel selected by the other tuner means.

  A program according to the present invention is a program for causing a computer to execute an operation of a specific image detection apparatus that detects a specific image from a still image included in a video composed of continuous frame-unit images. A candidate image detection step of calculating a luminance change ratio between a frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image A candidate image feature amount detecting step for detecting one or more feature amounts having different characteristics in the candidate image, a feature amount of the candidate image, and one or more feature amounts having different characteristics in the registered image stored in the storage unit A detection step of detecting the candidate image as a specific image when the similarity is equal to or greater than a predetermined threshold, and the specific image detected in the detection step The specific image detecting apparatus has at least two tuner means, and in the detecting step, the specific image is obtained from the video of the reception channel selected by one of the tuner means. In the presenting step, the video of the reception channel selected by the other tuner means is displayed on the first display means, and the specific image is displayed on the second display means.

  Furthermore, the program according to the present invention is a program for causing a computer to execute an operation of detecting a specific image from still images included in a video composed of continuous frame-unit images. A candidate image detection step of calculating a ratio of a luminance change between adjacent frame images that are temporally close to each other and detecting a frame image having a ratio of the luminance change equal to or less than a predetermined threshold as a candidate image, and the candidate image A candidate image feature quantity detecting step for detecting one or more feature quantities having different characteristics in the image, a similarity between the feature quantity of the candidate image and one or more feature quantities having different characteristics in the registered image stored in the storage unit A detection step for detecting the candidate image as a specific image when the threshold value is equal to or greater than a predetermined threshold; and the specific image detected in the detection step is displayed on the display means. To execute a that presentation step, in the present process, the specific image detected by said detecting step to keep multiple simultaneously displayed on the display means by reducing the specific image of the plurality of.

  A specific image detection system according to the present invention detects a specific image including a still image portion included in a video composed of images of successive frames, and displays the specific image on at least one display means. The ratio of the luminance change between at least one display means, storage means for storing one or more feature quantities having different characteristics in the registered image, and the frame image and the adjacent frame image temporally adjacent thereto And a candidate image detection unit that detects a frame image having a luminance change rate equal to or less than a predetermined threshold as a candidate image, and a feature amount detection unit that detects one or more feature amounts having different characteristics in the candidate image. Detecting means for detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and the feature amount of the registered image is equal to or greater than a predetermined threshold; A display unit that is connected to the display unit and displays the specific image detected by the detection unit on the at least one display unit; and a specific image detection unit that includes at least two tuner units. The means displays on the display means the video of the reception channel selected by one tuner means, and the detection means detects the specific image from the video of the reception channel selected by the other tuner means.

  In addition, the specific image detection system according to the present invention detects a specific image including a still image portion included in a video composed of continuous frame-unit images, and displays the specific image on at least one display means. A detection system, a luminance change between at least one display means, storage means for storing one or more feature quantities having different characteristics in a registered image, and a frame image and a close frame image temporally adjacent thereto A candidate image detecting means for detecting a frame image whose luminance change ratio is equal to or less than a predetermined threshold as a candidate image, and feature quantity detection for detecting one or more feature quantities having different characteristics in the candidate image And detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and the feature amount of the registered image is equal to or greater than a predetermined threshold value. And a specific image detection means comprising: a display means connected to the display means for displaying the specific image detected by the detection means on the at least one display means; and at least two tuner means. The detecting means detects the specific image from the video of the reception channel selected by one tuner means, and the presenting means is connected to at least the first and second display means, and the other tuner means The image of the selected reception channel is displayed on the first display means, and the specific image is displayed on the second display means.

  Furthermore, the specific image detection system according to the present invention detects a specific image including a still image portion included in a video composed of continuous frame-unit images, and displays the specific image on at least one display means. A detection system, a luminance change between at least one display means, storage means for storing one or more feature quantities having different characteristics in a registered image, and a frame image and a close frame image temporally adjacent thereto A candidate image detecting means for detecting a frame image whose luminance change ratio is equal to or less than a predetermined threshold as a candidate image, and feature quantity detection for detecting one or more feature quantities having different characteristics in the candidate image Means for detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and the feature amount of the registered image is equal to or greater than a predetermined threshold. Means, a display means connected to the display means and displaying the specific image detected by the detection means on the at least one display means, and a specific image detection means comprising at least two tuner means. The presenting means holds a plurality of the specific images detected by the detecting means, reduces the plurality of specific images, and displays them simultaneously on the display means.

  According to the present invention, an important image or the like specialized for the user is registered as a registered image in advance, and only a frame image whose luminance change rate is equal to or less than a predetermined threshold is selected from continuous images such as TV broadcasts. Is detected as a candidate image, and a candidate image similar to a registered image is detected as a specific image, so that a registered image can be selected with high efficiency and detected in real time. Since the image is specified by the feature amount after the candidate image is detected, the frequency of erroneous detection can be reduced.

It is a block diagram which shows the specific image detection system in embodiment of this invention. It is explanatory drawing explaining the flow of the data in the specific image detection apparatus in embodiment of this invention. It is a flowchart which shows the specific image detection method in the specific image detection apparatus in embodiment of this invention. It is a schematic diagram which shows the feature-value table of the specific image detection apparatus in embodiment of this invention. It is a flowchart which shows the feature-value acquisition method of the registration image in embodiment of this invention. It is a flowchart which shows the still determination method (detection method of a candidate image) in embodiment of this invention. It is a flowchart which shows the calculation method of the histogram in embodiment of this invention. (A) thru | or (c) is a figure which shows an example of the registration image which the user gave the registration instruction | indication, Comprising: It is a schematic diagram of the screen which respectively shows the forecast weather of East Japan, the forecast temperature of East Japan, and the weekly weather of East Japan. . (A) is a schematic diagram showing a state where a plurality of detected specific images are stored and simultaneously displayed on a sub-monitor, (b) is a display example on the main monitor at that time, and (c) is a detection screen (specific It is a schematic diagram which shows a screen.

  Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings. In this embodiment, the present invention is applied to a specific image detection system that detects and displays a specific image having important information designated by a user from consecutive images.

  In this specific image detection system, a user registers a registered image in advance, searches for an image included in a video such as a television (TV) broadcast, detects a specific image similar to the registered image, and displays this in real time. (Presentation, presentation). Here, the similarity determination with the registered image is performed, and the specific image is detected. Since the detection result is displayed at various timings, in this embodiment, the user evaluates the detection result. Means shall be provided. That is, even if the evaluation of the detection result changes depending on the state of the user at that time, the user's evaluation at that time with respect to the detection result can be fed back. Thereby, for example, the determination criterion in similarity determination can be changed and reflected in subsequent detection. In this way, important information (specific image) can be efficiently detected and displayed through interaction with the user.

  FIG. 1 is a block diagram showing a specific image detection system according to the present embodiment. As shown in FIG. 1, the specific image detection device 20 detects a specific image registered by a user from, for example, a TV broadcast video, and the user views a video such as a TV broadcast or a video. The main monitor 12 to be connected to the sub-monitor 11 for displaying the specific image detected from the video, and the specific image detection system 30 is configured. In this embodiment, in addition to the main monitor 12 such as a television receiver used when viewing normal TV video or the like, an efficient display can be achieved by using the sub-monitor 11 that displays the detected specific image. Do.

  In general, TV broadcasts have a plurality of receivable broadcasts, and each broadcast station broadcasts its own content, but it can often be said that they are not arranged in an order that is easy for the user to see. For example, taking a news program as an example, if a plurality of broadcasting stations are broadcasting a news program at the same time, the contents of the same news program are different for each broadcasting station. Therefore, there is a possibility that the target news may be overlooked. On the other hand, there may be a case where the user is watching news that is not personally interesting to the user. What is required here is a system that can efficiently view a specific part of a plurality of programs in order from the important part to the user. Other than news, for example, when you want to watch the weather forecast, you must watch news that you are not interested in because you do not know when it starts in the news program, or it is in a region other than the region where the user lives There are inconveniences such as going to see. In addition, when viewing other video supplied from an external device other than TV broadcast, if you want to view specific information of TV broadcast, you must temporarily stop viewing the other video. Inconvenient. The specific image detection system of the present invention eliminates such inconveniences and makes it possible to display a program efficiently.

  Here, in the present embodiment, the specific image detection device 20 has a plurality of tuner means, displays the video of the reception channel selected by one tuner means on the main monitor 12, and uses other tuner means. A specific image can be detected from the video of the selected reception channel and displayed on the sub-monitor 11. For example, when there are two tuner means, since there are a plurality of channels in TV broadcasting, a plurality of channels specified by the user or all channels are selected in a time division manner, and each channel selected in this time division is selected. A specific image can be detected from the TV video. Further, tuner means for the main monitor 12 and tuner means corresponding to all TV channels may be provided in order to detect a specific image.

  Further, tuner means may be provided in the sub monitor 11 and the main monitor 12 so that the sub monitor 11 and the main monitor 12 receive TV broadcasts and supply video to the specific image detection device 20. Alternatively, an image may be supplied by connecting an external device having tuner means to the specific image detection system. If the STD or the like such as a video deck, a DVD, or a BS (Broadcast Satelite) digital tuner is connected to the specific image detection device 20 as an external device, for example, various videos are supplied to the main monitor 12 and viewed. The specific image can be displayed on the sub monitor 11 or the main monitor 12 as necessary.

  Hereinafter, the specific image detection apparatus according to the present embodiment will be described in detail. FIG. 2 is an explanatory diagram for explaining the flow of data in the specific image detection apparatus of the present embodiment. As shown in FIG. 2, the specific image detection device 20 includes an image feature quantity acquisition unit (feature quantity detection unit) 3, a delay circuit (DL) 5, and a stillness determination unit (to which a video signal is supplied via the input terminal 1. (Candidate image detection means) 6, an image feature quantity acquisition section (feature quantity detection means) 7 to which a candidate image determined to be stationary from the stillness determination section 6 is supplied, and a registration instruction signal from the user via the input terminal 2 The feature amount of the input image (registered image) when supplied is supplied, the feature amount table (storage means) 4 for storing the feature amount, the features of the feature amount table 4 and the candidates supplied from the image feature amount acquisition unit 7 Similarity determined to be similar by the similarity determination unit 8 that is a detection unit that determines a similarity to the feature amount of the image and detects a candidate image similar to the registered image, and the similarity determination unit 8 An image is supplied to the sub monitor 11 and an input terminal 0 evaluation signal of the user is supplied on the displayed similar image on the sub monitor 11 through, and a similar image according to the voted presentation control unit supplies the main monitor 12 (presenting means) 9.

  Here, it is assumed that the video signal (input image sequence) is supplied via the input terminal. However, as described above, the supplied video signal is selected by providing the tuner means in the apparatus. It may be a TV image of the receiving channel or an image supplied from the outside.

  The delay circuit 5 supplies an input frame unit image to the stillness determination unit 6 with a predetermined frame delay, for example, one frame. The stillness determination unit 6 determines that the input frame image is a candidate image when the luminance change between the input frame image and a delay frame delayed by a predetermined frame from the input image supplied from the delay circuit 5 is equal to or less than a predetermined threshold. judge. Here, the candidate image is a still image or an image including a still image portion. An image that the user desires to detect is registered as a registered image. The threshold value is appropriately set depending on whether the registered image is a still image or an image including a still image portion.

  When the registration instruction signal is supplied via the terminal 2, the image feature amount acquisition unit 3 acquires a feature amount from the input image (registered image) at that time and sends it to the feature amount table 4. The feature amount table 4 stores the feature amount of a registered image that the user wants to register for reasons such as whether the user is important or necessary among input images.

  In addition, the image feature amount acquisition unit 7 acquires feature amounts from the candidate images, similarly to the image feature amount acquisition unit 3. In FIG. 2, the image detection apparatus 20 includes the image feature amount acquisition units 3 and 7, but the feature amount acquired by any one of the image feature amount acquisition units may be used.

  The similarity determination unit 8 compares all the feature amounts stored in the feature amount table 4 with the feature amounts detected from the candidate images, and determines the similarity. Specifically, for example, in order to determine the similarity of feature amounts using a predetermined threshold (determination threshold), the feature amount table 4 includes a determination threshold for each feature amount together with the feature amount of each registered image. It is remembered. The similarity determination unit 8 outputs a candidate image whose similarity is equal to or higher than a predetermined determination threshold value to the presentation control unit 9 as a similar image, and causes the sub monitor 11 to display the candidate image.

  The sub monitor 11 is a monitor that displays the detected similar image for the user to evaluate. The user's evaluation is input to the presentation control unit 9 via the terminal 10. For example, a plurality of user evaluations can be prepared, and the control of the presentation control unit 9 can be set based on the evaluations.

  When the user evaluation is input via the terminal 10, the presentation control unit 9 controls the image output to the sub-monitor 11 and the main monitor 12 based on the evaluation and rewrites data stored in the feature amount table 4. I do. For example, when the user views a similar image displayed on the sub-monitor 11 and this similar image is an image desired by the user (registered image), the similar image is identified by an evaluation signal from the user who evaluates the similar image. Can be output to the main monitor 12. On the other hand, when the similar image displayed on the sub-monitor 11 is different from the registered image, that is, when it is a false detection, the presentation control unit 9 uses the evaluation signal from the user pointing out that it is a false detection. It is possible to control to rewrite the determination threshold value stored in the feature amount table 4 at the time of detection. In this manner, by rewriting the determination threshold according to the user's evaluation, it is possible to prevent the unnecessary image and the image that is erroneously detected from being detected again. Furthermore, when an image similar to an unnecessary registered image is detected, for example, when it is a registered image that has been registered before, it can be set to be discarded.

  In addition, the specific image detection apparatus 20 according to the present embodiment outputs a similar image to the sub-monitor 11 and the specific image to the main monitor 12 by the presentation control unit 9, so that, for example, the user needs an image related to the weather forecast. However, by registering this, the specific image detection device 20 detects an image related to the weather forecast as a specific image. During that time, the user can watch a movie on the main monitor 12 by, for example, a video or the like. Other TV programs can be viewed, and a registered image can be detected from a so-called back program that is not viewed on the main monitor 12 and displayed on the sub-monitor 11.

  Next, Table 1 below shows examples of feature amounts stored in the feature amount table 4.

  In the present embodiment, a case where a luminance histogram and a stationary time are used from the feature amounts shown in Table 1 will be described. Here, for example, the specific image may be detected by using, for example, all the feature values shown in Table 1 above. However, it is robust to set a strict judgment criterion according to the input video, for example. The type of feature amount, the determination criterion (determination threshold), and the like are set as appropriate in consideration of cases that may be affected. In this embodiment, the reason for using the luminance histogram and the stationary time when detecting a specific image from TV broadcast video is to improve the robustness even in TV broadcast where noise is expected to enter. Because. In the following description, the specific image to be detected is a still image. However, as described above, an image including a still image portion, that is, a background portion, for example, is a moving image, and only a predetermined area is still. Even in such a case, it can be detected in the same manner.

  FIG. 3 is a flowchart showing a specific image detection method in the specific image detection apparatus of the present embodiment. As shown in FIG. 3A, first, stillness determination is performed to detect a candidate image (hereinafter referred to as a still image) from continuous images such as an input television broadcast video (step S1). ). On the other hand, when an image registration signal is input together with the input image by the user, the input image is registered as a registered image, so that the feature amount of the registered image is acquired (detected) (step S11). It is stored in the feature quantity table 4 (step S12). In the feature amount table 4, the feature amount of the registered image designated by the user, that is, in this embodiment, whenever the registration is instructed by the user, the luminance histogram and the still time of the registered image are acquired. Is stored and saved as needed.

  On the other hand, when a still image is detected by performing still determination from the input image, the feature amount of the still image is detected, and the obtained feature amount is sent to the similarity determination unit 8 and stored in the feature amount table 4. It compares with the feature-value of all the registration images which exist, and similarity determination is performed based on a threshold value (step S2). In the present embodiment, similarity determination is performed based on a correlation coefficient e, which will be described later, with respect to a luminance histogram, a ratio of unchanged blocks without change, and a still time that is a count value of a still counter (not shown).

  FIG. 4 is a schematic diagram showing the feature amount table 4. As shown in FIG. 4, the feature amount table 4 is assigned an identification number (No. 0, 1, 2,...) For each registered image, and a luminance histogram, a stationary time, and a sequence corresponding to each registered image. Determination thresholds relating to correlation coefficients, ratios of changed blocks, time errors, and the like are stored.

  Returning to FIG. 3, when it is determined in step S <b> 3 that the registered image and the still image are similar, the data of this image is sent to the presentation control unit 9, and a display instruction is sent to the sub-monitor 11 for display. (Step S3). In the present embodiment, after the still image is displayed, for example, an alarm for prompting the user's attention is sounded, and a standby state is waited for an evaluation input from the user via the input terminal 10 (steps). S4). On the other hand, if it is determined that they are not similar, processing proceeds to the next frame.

  Next, the user evaluates the displayed image. As described above, various user evaluations are conceivable. A plurality of evaluations can be set, and processing corresponding to the evaluations can be set. In this embodiment, when it is determined that the displayed image is a registered image specified by the user (determined that it is necessary) (Good), and it is determined that it is not a registered image specified by the user (unnecessary) A case where two patterns of “Nogood” are determined will be described. Here, when the evaluation of Nogood is input, the value of the determination threshold value of the registered image in the feature amount table 4 used when it is determined to be similar is changed (updated) (step S6), and the next frame Proceed to the process. On the other hand, if it is evaluated as Good and it is determined that the similar image is a registered image, the specific image is output to the main monitor 12 (step S7), and the process proceeds to the next frame. As described above, when the user's evaluation is Nogood, by rewriting the value of the determination threshold value and tightening the criteria for determining similarity, it is possible to prevent the same erroneous detection from occurring in subsequent detections. In this example, the specific image is output to the main monitor 12. However, the specific image is not output to the main monitor 12. For example, a storage unit is provided in the sub-monitor 11 and is held for a predetermined period. You may make it display on.

  Hereinafter, the specific image detection method in the present embodiment will be described in more detail. First, a feature amount acquisition method for registered images in the image feature amount acquisition unit 3 will be described. FIG. 5 is a flowchart illustrating a method for acquiring a feature amount of a registered image. In FIG. 5, when a pixel value (F (i, pxl)) of a frame (hereinafter referred to as i frame) image is input and a registration instruction signal is input from the user via the terminal 2, that is, an i frame image. A case will be described in which an instruction for acquiring a feature amount is issued and an i-frame image is registered as a registered image. In the present embodiment, as described above, the luminance histogram and the stationary time are acquired as the feature amount of the registered image and stored in the feature amount memory 4.

  As shown in FIG. 5, first, when the i-frame image data is supplied via the terminal 1 and the registration instruction signal is supplied via the terminal 2, and the i-frame is selected as the registered image, the image feature amount The acquisition unit 3 divides this i-frame image data into prescribed blocks, obtains a histogram of luminance values in each block, and holds the obtained luminance histogram H (i, j) (step S21). Here, H (i, j) represents the luminance histogram of the jth block in the i frame. Further, the obtained luminance histogram is copied as a feature value and stored in the feature value table 4 (step S22).

  Next, the frame number is incremented (step S23), and the pixel value (F (i + 1, pxl)) of the (i + 1) frame that is temporally continuous with the i frame is input. Similarly to the i frame, the image feature quantity acquisition unit 3 obtains and holds a luminance histogram H (i + 1, j) in units of blocks in the (i + 1) frame image.

  A correlation coefficient e between spatially corresponding blocks between the obtained luminance histogram H (i + 1, j) and the previously obtained luminance histogram H (i, j) is expressed by the following equation (1). calculate.

  Then, the calculated correlation coefficient e is compared with a predetermined threshold value (Th_coef) (step S26). If the correlation coefficient e (j) in the j-th block is equal to or less than the threshold (Th_coef), this is counted as a change block. If the counted change block ratio e2 (j) is less than the threshold (Th_diff), a still image As a result, the count value (stillcnt) of the still counter is incremented (step S27), the process proceeds to the next frame ((i + 2) frame), and the processing from step S23 to step S27 is repeated. On the other hand, if the change block ratio e2 (j) is equal to or greater than the threshold (Th_diff), the count value of the stationary counter is stored in the feature amount table 4 (step S28), and the process ends. Here, the count value of the still counter is the still time of the registered image acquired by the image feature amount acquisition unit 3. In this way, the image feature quantity acquisition unit 3 can acquire the luminance histogram and the stationary time as the feature quantity of the i-frame image designated as the registered image and store it in the feature quantity table 4.

  Next, a still image detection method of the stillness determination unit will be described. As for the still image, the delayed frame image and the input frame image that have passed through the frame delay circuit 5 are sent to the stillness determination unit 6, and whether or not the still image is still is determined from the ratio of the changed blocks in the same manner as in step S 27 described above. Determination is made and a still image (candidate image) is detected. FIG. 6 is a flowchart showing a still image detection method in the stillness determination unit 6.

  The stillness determination unit 6 divides the i-frame image input via the terminal 1 into prescribed blocks, obtains a histogram of luminance values in each block, and holds the obtained luminance histogram H (i, j). (Step S31). Next, for example, the frame number is incremented by 1 (step S32), and the pixel value (F (i + 1, pxl)) of (i + 1) frame is input (step S33). Also, the luminance histogram H (i + 1, j) in block units is obtained and held (step S34).

  Then, between the obtained luminance histogram H (i + 1, j) and the previously obtained luminance histogram H (i, j), the correlation coefficient e (j) between spatially corresponding blocks is expressed by the above formula ( 1), the correlation coefficient e (j) is compared with a predetermined threshold (Th_coef) (step S35), and the ratio of changed blocks is detected (step S36). If the correlation coefficient e (j) of the j-th block is equal to or less than the threshold (Th_coef), this is counted as a change block. If the counted change block ratio e2 (j) is less than the threshold (Th_diff), i The brightness histogram of the frame image and the counter value of the still counter are stored (step 37), and this i-frame image is regarded as a still image and sent to the image feature quantity acquisition unit 7. On the other hand, if the ratio of changed blocks is greater than or equal to the threshold value (Th_diff), the process ends and the process proceeds to the next frame.

  The i-frame image that is regarded as a still image and sent to the image feature amount acquisition unit 7 is counted and acquired as a feature amount in the same manner as the acquisition of the still time of the feature amount acquisition unit 3 shown in FIG. The

  In FIG. 6, for simplification of explanation, the correlation coefficient of the corresponding block is compared between the i frame and the (i + 1) frame. For example, in the example illustrated in FIG. 2, the stillness determination unit 6 includes an input frame image (i frame image), and a correlation between adjacent frames. A frame image ((i-2) frame image) delayed by a predetermined frame from the input frame image is supplied via the delay circuit 5, and the correlation between the two is determined.

  Next, a histogram calculation method in the image feature amount acquisition units 3 and 7 will be described. In the present embodiment, as described above, the luminance histogram is used as the stillness determination and the feature amount of the stillness determination unit 6. The histogram generation unit for calculating the luminance histogram may be separately prepared in, for example, the image feature amount acquisition units 3 and 7 and the stillness determination unit 6, or a luminance histogram generation unit (not shown in FIG. 2) is prepared, If necessary, image data may be sent to the luminance histogram generation unit to generate a luminance histogram.

  FIG. 7 is a flowchart showing a histogram calculation and a method for comparing this histogram between frames. Here, a case where the luminance histograms between the i frame and the (i-2) frame are compared will be described. As shown in FIG. 7, first, an input frame image (referred to as an i-frame image) is divided into blocks of a prescribed size (step S41). For example, the frame image can be 720 × 480 pixels, for example, and the block size can be 32 × 32 pixels, for example. Then, the luminance histogram H (i, j) of the jth block divided in the i frame is calculated. Next, the already calculated jth luminance histogram H (i-2, j) of the (i-2) frame is read (step S43), and the luminance of the jth block in the i frame is calculated by the above equation (1). A correlation coefficient e (i, j) of the histogram is calculated (step S45), and it is determined whether or not the block is a change block by threshold determination. At this time, if it is determined that the block is a change block, 1 is set in the flag E (i, j), and if it is determined that the block is a non-change block, 0 is set in the flag E (i, j). . The luminance histogram H (i, j) is saved for use when comparing the correlation coefficient with the (i + 2) frame, the counter j is incremented (step S46), and whether or not all the locks have been completed. If it is determined (step S47) that the change block determination has not been completed for all locks, the processing from step S42 to step S47 is repeated, and if the change block determination has been made for all blocks, the process ends. . Thus, the value of the obtained all-lock flag E (i, j) is added, and if this value is equal to or less than a predetermined value, the i-frame image can be determined as a still image.

Hereinafter, as a specific example of the present embodiment, a case where a weather forecast program is a registered image and this is detected will be described. The following are the features of the weather forecast program. That is,
1. 1. It is configured by combining similar screens each time. 3. Broadcast at almost the same time every time, 3. Many are not set as program frames, and the exact start time is unknown from the program guide. 4. Frequency of using still images is high. It has features such as variations in screens that are important for each user, such as regional dependency.

  In such a weather forecast program, similar images are used repeatedly, so that a registered image (registration screen) can be used for a long time. In addition, although it is possible to know the approximate start time of a weather forecast program, it is difficult to accurately audition only the weather forecast program in situations such as being included in a news program. When viewing the content, it is extremely inconvenient because the user is not interested in the content. Furthermore, since there are contents that become unnecessary information depending on the area where the user lives, the advantage of selectively extracting images is great.

  The user issues a registration instruction if there is a desired image at the time of program preview. FIGS. 8A to 8C are diagrams showing examples of registration images for which the user has issued a registration instruction, and are schematic diagrams of screens showing predicted weather in East Japan, predicted temperature in East Japan, and weekly weather in East Japan, respectively. It is.

  Here, a case where the user registers information related to the Kanto region as important information will be described. First, when an image similar to the registered image is detected, it is displayed as a detection result on the sub-screen. 9A is a schematic diagram showing a similar image displayed on the sub screen (detection screen), FIG. 9B is a display example on the main monitor 12 at that time, and FIG. 9C is a detection. It is a schematic diagram which shows a mode that the several specific image made was preserve | saved and displayed on the sub monitor 11 simultaneously. For example, as shown in FIG. 9B, when the user is listening to a video such as a movie on which the weather forecast program is not displayed on the main monitor 12, as shown in FIG. When an image similar to the registered image is broadcast on another station, the similar image of the registered image is detected and displayed on the sub-monitor 11.

  Here, the user determines whether or not to audition the continuation of the detection result. For example, when the user gives an instruction to audition, the detection image displayed on the sub monitor 11 is switched and displayed on the main monitor 12. It is possible to make it. Here, it is possible to prevent an unpleasant display from being made depending on the situation of the place by entrusting the determination to the user. In addition, when a screen different from the registered image is detected and the user has not selected trial listening, the determination threshold value of the registered screen corresponding to the detection result is changed to prevent subsequent erroneous detection. It is possible to improve the detection efficiency and generate a detection determination threshold with high accuracy inherent to the screen.

  Furthermore, in the present embodiment, a plurality of sub-monitors 11 may be prepared, and the detected images may be saved, reduced, and displayed as a past list as shown in FIG. 9C. Considering the creator's intention, still images often contain important information that the viewer wants to pay attention to. Furthermore, this is an image devised to express some information by one piece. For this reason, a still image has a larger amount of information than one frame in a normal moving image, and a large amount of information can be transmitted only by displaying one frame. Therefore, by saving, displaying for a certain period of time, and displaying a reduced list, etc., the user can look back later to check the information, and it is extremely convenient when organizing the information. .

  Further, by preparing a plurality of sub-monitors 11 having tuner means, using the tuners of each sub-monitor 11, and supplying video signals from each sub-monitor 11, the number of broadcast stations that can be searched simultaneously can be increased.

  An example of an effect similar to a weather forecast program is a news title screen of a news program. In news programs, an announcer's anchor shot (bust shot) and news title telop are often displayed for each news item, and there is almost no change for a certain period of time. Also, depending on the news program, the effect and background of the telop are constant, which is suitable for the detection target. In addition, news programs may not always be broadcast as scheduled in the program guide. By registering such images, you can view news programs from different broadcast stations as detection results while viewing any program. It can be viewed only if it is of interest to the user or if the content is of interest. Examples of registered images suitable for such registration include market information, sports game results, and a title screen of a specific corner in a program.

  In the present embodiment, an important image specialized for the user is selected in advance from continuous images such as TV broadcasts, and a similar image of this image is detected in real time and notified to the user in a more efficient manner. Important information can be displayed. In addition, by detecting specific images from multiple broadcasts by having multiple tuner means or searching multiple channels in a time-sharing manner, it is possible to monitor back programs that are different from the program that the user is currently viewing It becomes.

  Further, by determining still images, candidate still images are narrowed down, and by determining the similarity between the still image and the registered image, the amount of data for acquiring the feature amount and determining the similarity may be small. It is possible to detect a specific image in real time from a video such as the above, and it is possible to efficiently select and view important information for the user and to prevent oversight of important information.

  Moreover, the frequency of erroneous detection can be reduced by using a plurality of feature amounts in an integrated manner. Furthermore, the accuracy of image detection, which is difficult, can be set flexibly by displaying the detection result for a certain period of time and prompting evaluation from the user. Furthermore, by making the accuracy flexible, the robustness can be improved, and the user's judgment in the displayed state can be utilized.

  1, 2, 10 input terminals, 3 image feature amount acquisition unit, 4 feature amount table, 5 delay circuit, 6 stillness determination unit, 7 image feature amount acquisition unit, 8 similarity determination unit, 9 presentation control unit, 11 sub monitor, 12 main monitor, 20 specific image detection device

Claims (13)

A specific image detection device for detecting a specific image including a still image portion included in a video composed of images in consecutive frames,
Storage means for storing one or more feature quantities having different characteristics in the registered image;
Candidate image detecting means for calculating a ratio of luminance change between the frame image and a neighboring frame image close in time to the frame image, and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
Feature quantity detection means for detecting one or more feature quantities having different characteristics in the candidate image;
Detecting means for detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and the feature amount of the registered image is equal to or greater than a predetermined threshold;
Presenting means for displaying on the display means the specific image detected by the detecting means;
Having at least two tuner means;
The presenting means displays on the display means the video of the receiving channel selected by one tuner means,
The specific image detection device, wherein the detection means detects the specific image from the video of the reception channel selected by the other tuner means. A specific image detection device for detecting a specific image including a still image portion included in a video composed of images in consecutive frames,
Storage means for storing one or more feature quantities having different characteristics in the registered image;
Candidate image detecting means for calculating a ratio of luminance change between the frame image and a neighboring frame image close in time to the frame image, and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
Feature quantity detection means for detecting one or more feature quantities having different characteristics in the candidate image;
Detecting means for detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and the feature amount of the registered image is equal to or greater than a predetermined threshold;
Presenting means for displaying on the display means the specific image detected by the detecting means;
Having at least two tuner means;
The detection means detects the specific image from the video of the reception channel selected by one tuner means,
The presenting means is connected to the first display means and the second display means, displays the image of the reception channel selected by the other tuner means on the first display means, and displays the specific image on the second display means. A specific image detection device for displaying on the display means. The specific image detection apparatus according to claim 2, wherein the detection unit detects the specific image from videos of a plurality of reception channels selected in a time division manner. A specific image detection device for detecting a specific image including a still image portion included in a video composed of images in consecutive frames,
Storage means for storing one or more feature quantities having different characteristics in the registered image;
Candidate image detecting means for calculating a ratio of luminance change between the frame image and a neighboring frame image close in time to the frame image, and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
Feature quantity detection means for detecting one or more feature quantities having different characteristics in the candidate image;
Detecting means for detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and the feature amount of the registered image is equal to or greater than a predetermined threshold;
Display means for displaying the specific image detected by the detection means on the display means,
The presenting means holds a plurality of the specific images detected by the detection means, reduces the plurality of specific images, and simultaneously displays the specific images on the display means. A specific image detection method of a specific image detection device for detecting a specific image including a still image portion included in a video composed of images of successive frames,
On the computer,
A candidate image detecting step of calculating a luminance change ratio between the frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
A candidate image feature amount detecting step of detecting one or more feature amounts having different characteristics in the candidate image;
A detection step of detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and one or more feature amounts having different characteristics in the registered image stored in the storage unit is equal to or greater than a predetermined threshold; ,
A display step of displaying the specific image detected in the detection step on a display means;
The specific image detection apparatus has at least two tuner means,
In the presenting step, the video of the reception channel selected by one tuner means is displayed on the display means,
In the detection step, a specific image detection method for detecting the specific image from a video of a reception channel selected by the other tuner means. A specific image detection method of a specific image detection device for detecting a specific image including a still image portion included in a video composed of images of successive frames,
On the computer,
A candidate image detecting step of calculating a luminance change ratio between the frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
A candidate image feature amount detecting step of detecting one or more feature amounts having different characteristics in the candidate image;
A detection step of detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and one or more feature amounts having different characteristics in the registered image stored in the storage unit is equal to or greater than a predetermined threshold; ,
A display step of displaying the specific image detected in the detection step on a display means;
The specific image detection apparatus has at least two tuner means,
In the detection step, the specific image is detected from the video of the reception channel selected by one tuner means,
In the presenting step, a specific image detecting method in which an image of a reception channel selected by the other tuner unit is displayed on a first display unit, and the specific image is displayed on a second display unit. A specific image detection method of a specific image detection device for detecting a specific image including a still image portion included in a video composed of images of successive frames,
On the computer,
A candidate image detecting step of calculating a luminance change ratio between the frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
A candidate image feature amount detecting step of detecting one or more feature amounts having different characteristics in the candidate image;
A detection step of detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and one or more feature amounts having different characteristics in the registered image stored in the storage unit is equal to or greater than a predetermined threshold; ,
A display step of displaying the specific image detected in the detection step on a display means;
In the presenting step, a plurality of the specific images detected in the detection step are held, and the specific images are reduced and simultaneously displayed on the display unit. A program for causing a computer to execute an operation of a specific image detection device that detects a specific image from a still image included in a video composed of continuous frame-unit images,
On the computer,
A candidate image detecting step of calculating a luminance change ratio between the frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
A candidate image feature amount detecting step of detecting one or more feature amounts having different characteristics in the candidate image;
A detection step of detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and one or more feature amounts having different characteristics in the registered image stored in the storage unit is equal to or greater than a predetermined threshold; ,
A display step of displaying the specific image detected in the detection step on a display means;
The specific image detection apparatus has at least two tuner means,
In the presenting step, the video of the reception channel selected by one tuner means is displayed on the display means,
In the detection step, a program for detecting the specific image from the video of the reception channel selected by the other tuner means. A program for causing a computer to execute an operation of a specific image detection device that detects a specific image from a still image included in a video composed of continuous frame-unit images,
On the computer,
A candidate image detecting step of calculating a luminance change ratio between the frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
A candidate image feature amount detecting step of detecting one or more feature amounts having different characteristics in the candidate image;
A detection step of detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and one or more feature amounts having different characteristics in the registered image stored in the storage unit is equal to or greater than a predetermined threshold; ,
A display step of displaying the specific image detected in the detection step on a display means;
The specific image detection apparatus has at least two tuner means,
In the detection step, the specific image is detected from the video of the reception channel selected by one tuner means,
In the presenting step, a program for displaying an image of a reception channel selected by the other tuner means on a first display means and displaying the specific image on a second display means. A program for causing a computer to execute an operation of detecting a specific image from a still image included in a video composed of continuous frame-unit images,
On the computer,
A candidate image detecting step of calculating a luminance change ratio between the frame image and a close frame image that is temporally adjacent thereto and detecting a frame image having the luminance change ratio equal to or less than a predetermined threshold as a candidate image; ,
A candidate image feature amount detecting step of detecting one or more feature amounts having different characteristics in the candidate image;
A detection step of detecting the candidate image as a specific image when the similarity between the feature amount of the candidate image and one or more feature amounts having different characteristics in the registered image stored in the storage unit is equal to or greater than a predetermined threshold; ,
A display step of displaying the specific image detected in the detection step on a display means;
In the presenting step, a program that holds a plurality of the specific images detected in the detection step, reduces the plurality of specific images, and displays them simultaneously on the display means. A specific image detection system for detecting a specific image including a still image portion included in a video composed of continuous frame-unit images and displaying the specific image on at least one display means,
At least one display means;
The ratio of the luminance change between the storage means for storing one or more feature quantities having different characteristics in the registered image and the frame image and the adjacent frame image close in time to the frame image is calculated, and the luminance change ratio is predetermined. A candidate image detecting means for detecting a frame image that is equal to or less than a threshold value as a candidate image, a feature quantity detecting means for detecting one or more feature quantities having different characteristics in the candidate image, a feature quantity of the candidate image, and the registration A detection means for detecting the candidate image as a specific image when the similarity with the feature quantity of the image is equal to or greater than a predetermined threshold; and the specific image detected by the detection means connected to the display means. A display means for displaying on one display means, and a specific image detection means comprising at least two tuner means,
The presenting means displays on the display means the video of the receiving channel selected by one tuner means,
The specific image detection system, wherein the detection means detects the specific image from the video of the reception channel selected by the other tuner means. A specific image detection system for detecting a specific image including a still image portion included in a video composed of continuous frame-unit images and displaying the specific image on at least one display means,
At least one display means;
The ratio of the luminance change between the storage means for storing one or more feature quantities having different characteristics in the registered image and the frame image and the adjacent frame image close in time to the frame image is calculated, and the luminance change ratio is predetermined. A candidate image detecting means for detecting a frame image that is equal to or less than a threshold value as a candidate image, a feature quantity detecting means for detecting one or more feature quantities having different characteristics in the candidate image, a feature quantity of the candidate image, and the registration A detection means for detecting the candidate image as a specific image when the similarity with the feature quantity of the image is equal to or greater than a predetermined threshold; and the specific image detected by the detection means connected to the display means. A display means for displaying on one display means, and a specific image detection means comprising at least two tuner means,
The detection means detects the specific image from the video of the reception channel selected by one tuner means,
The presenting means is connected to at least the first and second display means, displays the video of the reception channel selected by the other tuner means on the first display means, and displays the specific image on the second display means. A specific image detection system to be displayed on the display means. A specific image detection system for detecting a specific image including a still image portion included in a video composed of continuous frame-unit images and displaying the specific image on at least one display means,
At least one display means;
The ratio of the luminance change between the storage means for storing one or more feature quantities having different characteristics in the registered image and the frame image and the adjacent frame image close in time to the frame image is calculated, and the luminance change ratio is predetermined. A candidate image detecting means for detecting a frame image that is equal to or less than a threshold value as a candidate image, a feature quantity detecting means for detecting one or more feature quantities having different characteristics in the candidate image, a feature quantity of the candidate image, and the registration A detection means for detecting the candidate image as a specific image when the similarity with the feature quantity of the image is equal to or greater than a predetermined threshold; and the specific image detected by the detection means connected to the display means. A display means for displaying on one display means, and a specific image detection means comprising at least two tuner means,
The presenting means holds a plurality of the specific images detected by the detection means, reduces the plurality of specific images, and simultaneously displays the specific images on the display means.

Images