JP2009296113A - Image display device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor capable of performing image correction which suppresses variation in the image quality, while the real time property is maintained. <P>SOLUTION: A hue information analysis portion 9 detects a hue information value Ci in each frame from a hue information H included in an image signal Db, and outputs it to a content feature determination portion 10. The content feature determination portion 10 determines features of a content in the image signal Db for one frame, based on the hue information value Ci. Moreover, the content feature determination portion 10 analyzes a video content for a plurality of frames and outputs a determined value Fi for a plurality of content features, the analysis result, to a correction control portion 4. The correction control portion 4 computes a correction parameter Pi, based on information Fi for a plurality of content features. A correction execution portion 5 uses the inputted correction parameter Pi to perform image correction processing such as tone correction for the image signal Db for one frame. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an image display device such as a television receiver.

  As a conventional image quality control device, for example, there is a television receiver disclosed in Patent Document 1. In the television receiver described in Patent Document 1, the name of the program being received, the genre of the program, the broadcast channel, the broadcast time zone, EPG (Electronic Program Guide) data, program information transmitted by data broadcasting, and character multiplex broadcasting The type of program can be acquired from any information in the program table of contents information to be transmitted, and the image quality can be automatically restored to the optimum image quality setting state desired by the viewer corresponding to the acquired type.

  Further, for example, Patent Document 2 discloses a method of determining and outputting a content genre while paying attention to video in multimedia data. In the content genre determination method described in Patent Document 2, the content type to which the multimedia data belongs is estimated from the history of the time interval when the screen is switched. Also, multimedia data whose content type is known in advance is compared with multimedia data including the input video, and the content type to which the input multimedia data belongs is estimated.

  Further, for example, Patent Document 3 discloses an image display device that can determine the characteristics of a broadcast video signal not only based on information transmitted from a broadcasting station but also based on a histogram in a luminance signal and determine video by genre. It is disclosed.

JP 2002-158941 A JP 11-261909 A JP 2007-124453 A

  In Patent Document 1, program information or EPG data is sent simultaneously with video information in the case of digital broadcasting such as CS broadcasting. In other words, these pieces of content information including genre information are limited in that they are information that is not included in conventional analog broadcasts or recorded videos such as DVDs.

  In addition, when the image quality setting is performed based on the genre of the transmitted video, it is possible that an inappropriate setting is performed. For example, when a movie with the theme of sports is received, the genre is determined to be a movie according to EPG data sent from a broadcasting station. However, originally, the optimum image quality setting state should be estimated from the video information. In this example, there is a problem that the optimum image quality setting for making sports look vivid cannot be realized.

  In Patent Document 2, in order to determine the content genre, it is necessary to hold a history of a specific time interval or multimedia data whose content type is known in advance, and multimedia data whose content type is known And the newly input multimedia data are compared to determine the content type of the newly input multimedia data. For this reason, there is a problem that the content genre cannot be determined for first-look data that does not hold multimedia data or data whose content type is not known.

  In consideration of the above-mentioned problems in Patent Document 1 or Patent Document 2, as described in Patent Document 3, a video image is extracted from a luminance histogram of an input video signal, and image processing adapted to the characteristics of the video signal is performed. The method of performing is considered. However, even in this method, the feature amount obtained from the image signal for one frame is not easily determined even in the same genre, and there is a problem that a defect may occur when image processing is performed. Further, although it is possible to calculate the feature amount of the video using the luminance histogram, there is a problem that the luminance histogram is insufficient to calculate more detailed features.

  The present invention has been made to solve the above-described problems, and an object thereof is to obtain an image processing apparatus capable of performing image correction while suppressing a change in image quality while maintaining real-time characteristics.

  According to a first aspect of the present invention, there is provided an image display device which receives an image signal and performs analysis processing in a plurality of divided regions divided in a color space on the color information obtained from the image signal for one frame. A tint information analysis unit that obtains a tint information value, the tint information including at least one of saturation information and hue information, and based on the tint information value, content information for the image signal A content feature determination unit that performs feature determination and obtains a determined content feature determination value, and an image correction unit that performs image correction on an image signal for one frame based on the determined content feature determination value.

  According to a sixth aspect of the present invention, the image display device according to the present invention receives an image signal and performs analysis processing in a plurality of divided regions divided in a color space on the color information obtained from the image signal for one frame. A hue information analysis unit that obtains a hue information value, and the hue information includes at least one of saturation information, hue information, and lightness information, and based on the hue information value, the image signal A content feature determination unit that determines a feature of the content and obtains a determined content feature determination value; and an image correction unit that performs image correction on an image signal for one frame based on the determined content feature determination value. The plurality of divided areas include at least a first and a second plurality of divided areas having different division contents, and the color information analyzing unit obtains a previous frame from the image signal for one frame. A color information detecting unit that detects color information; and a divided statistical processing unit that performs a divided statistical process on the color information in the first plurality of divided regions and outputs a first statistical result, The division statistical processing unit receives the division content instruction information, and when the division content instruction information instructs detailed division, performs a division statistical process on the color information in the second plurality of division areas and performs a second statistical result. Is further output, the color information value is calculated based on at least one of the first and second statistical results obtained by the divided statistical processing unit, and the divided based on the first statistical result And further including a divided region analysis processing unit that outputs content instruction information, wherein the divided region analysis processing unit, when the color information value obtained based on the first statistical result satisfies a predetermined requirement, And outputs the divided content instruction information for instructing the division.

  The content feature determination unit in the image display device according to claim 1 obtains the determined content feature determination value based on the color information value. The color information value is obtained from the color information analysis unit by analyzing the color information including at least one of saturation information and hue information in a plurality of divided areas divided in the color space. Value.

  Therefore, the image display device according to claim 1 performs image correction on the image signal for one frame by the image correction unit based on the determined content feature determination value that accurately reflects the characteristics of the content obtained from the image signal. As a result, it is possible to perform accurate image correction while adapting to the feature of the content or the process of changing the feature and suppressing the deterioration of the image quality while maintaining the real-time property.

  In the image display device according to claim 6 of the present invention, the division statistical processing unit of the color information analyzing unit obtains a first statistical result by performing division statistical processing on the color information in the first plurality of divided regions. ing.

  On the other hand, when the color information value obtained based on the first statistical result satisfies a predetermined requirement, the divided region analysis processing unit outputs divided content instruction information for instructing detailed division. When the division content instruction information indicates detailed division, the division statistical processing unit performs division statistical processing on the color information in the second plurality of divided areas and further obtains a second statistical result.

  Therefore, the image display device according to claim 6 has a relatively high-speed image correction process based on only the color information value based on the first statistical result, and a relatively high accuracy based on the color information value based on at least the second statistical result. Image correction processing with high image quality can be selectively performed based on the content of the image signal, so that more appropriate image correction can be performed according to the state of the image signal.

<Embodiment 1>
FIG. 1 is a block diagram showing a configuration of an image display apparatus according to Embodiment 1 of the present invention. As shown in the figure, the image display device according to the first embodiment includes an input terminal 1, a receiving unit 2, an image processing device 7, and a display unit 8.

  The input terminal 1 receives an image signal Da of a predetermined format used in a television or a computer. The receiving unit 2 receives the image signal Da via the input terminal 1, converts the image signal Da into a format that can be processed by the image processing device 7, and outputs the image signal Db. For example, the receiving unit 2 converts the image signal Da into several digital format image signals such as hue information H, saturation information S, and brightness information V, and then outputs the image signal Db. The receiving unit 2 that performs the above-described operation is configured by an A / D converter or the like when the image signal Da is an analog signal, and in the format when the input image signal Da is a digital signal. Consists of a predetermined demodulator and the like.

  The image processing apparatus 7 includes a color information analysis unit 9, a content feature determination unit 10, and an image correction unit 6, and the image correction unit 6 includes a correction control unit 4 and a correction execution unit 5. The image signal Db obtained from the reception unit 2 is input to the color information analysis unit 9 and the correction execution unit 5 of the image processing device 7, respectively.

  The hue information analysis unit 9 detects the hue information value Ci in each frame from the hue information H included in the image signal Db, and outputs it to the content feature determination unit 10.

  The content feature determination unit 10 determines the feature of the video content with respect to the image signal Db for one frame based on the color information value Ci, further analyzes the video content for a plurality of frames, and shows the analysis result. The multiple content feature determination value Fi (determined content feature determination value Fi) is output to the correction control unit 4.

  The correction control unit 4 calculates a correction parameter Pi used when the correction execution unit 5 performs image correction on the image signal Db based on the multiple content feature information Fi, and outputs the correction parameter Pi to the correction execution unit 5.

  On the other hand, the correction execution unit 5 performs image correction processing such as gradation correction on the image signal Db for one frame using the input correction parameter Pi, and displays the signal after image correction as the image signal Dc. Output to unit 8. The display unit 8 displays an image based on the input image signal Dc. The display unit 8 is, for example, a liquid crystal display, a DMD (Digital Micromirror Device) display, an EL (ElectoroLuminescence) display, a plasma display, a CRT (Cathode-Ray Tube) display, and a reflective type, a transmissive type, or a self-luminous device. Any display means can be applied.

  FIG. 2 is a block diagram showing details of one internal configuration example of the color information analysis unit 9 shown in FIG. As shown in FIG. 2, the tint information analysis unit 9 in the image display apparatus according to the first embodiment includes a hue detection unit 90c, a color space region division unit 91c, and a hue information analysis unit 92c.

  In the example illustrated in FIG. 2, the image signal Db obtained from the receiving unit 2 and including the hue information H, the saturation information S, and the lightness information V is input to the hue detection unit 90c.

  The hue detection unit 90c extracts the hue information H from the image signal Db for one frame and outputs it to the color space region division unit 91c.

  The color space region dividing unit 91c divides the color space into a plurality of divided (color) regions based on the hue information H in the image signal Db for one frame, performs division statistical processing in the divided regions, and performs region division information JDd ( H, D (frequency)) is detected and output to the hue information analysis unit 92c. In the first embodiment, the HSV color space is used as an example of the color space.

  The hue information analysis unit 92c calculates a hue signal feature value Ch in the image signal Db for one frame based on the area division information JDd generated by the color space area division unit 91c, and sets the hue information value Ci as the hue information value Ci. The color information value Ci is output to the content feature determination unit 10.

  In the configuration example of the color information analyzing unit 9 shown in FIG. 2, when the input video signal Db is an interlaced signal, the image is converted by the hue detecting unit 90c using two fields as an image signal for one frame. The area division information JDd is obtained from the color space area division unit 91c. In this manner, the hue detection unit 90c may perform preprocessing for obtaining the region division information JDd.

  Another example of pre-processing performed by the hue detection unit 90c is known when the video signal Db is displayed in a color system different from the hue information H, saturation information S, and lightness information V, such as an RGB signal. Conversion to hue information H, saturation information S, and lightness information V is also conceivable.

  FIG. 3 is generated by the color space region dividing unit 91c. It is explanatory drawing which showed an example of area | region division information JDd. In the example shown in FIG. 3, a hue histogram related to the hue information H is generated as the region division information JDd. Hereinafter, information defining the histogram of the hue information H in the image signal Db generated by the color space region dividing unit 91c is referred to as a hue histogram Dbh.

  In the example shown in FIG. 3, the horizontal axis in the figure indicates the hue angle, and the vertical axis indicates the frequency, that is, the number of pixels with respect to the hue angle of the image signal Db for one frame. In the following description, the hue signal of the image signal Db takes a value from “0” to “360”, but “0” to “256”, “0” to “1”, etc. Other scales may be used. In this example, 0 degrees represents red, followed by 60 degrees yellow, 120 degrees green, 180 degrees light blue, 240 degrees blue, 300 degrees purple, and again represents red at 360 degrees. .

  The hue histogram Dbh generated by the color space region dividing unit 91c according to the first embodiment divides, for example, 360 hue angles into six regions every 60 degrees, and the six regions are hues of the hue histogram. It is a horn. Then, a value near the central value at each hue angle, in this example, an integer value closest to the central value and larger than that is set as the representative value of the hue angle. For example, in a class composed of hue angles “0” to “60”, the central value is “30”, so the representative value of the hue angle is “30”. The numbers on the horizontal axis in FIG. 3 indicate the representative values of the hue angles.

  If the central value of the hue angle is an integer, the central value may be used as the representative value of the hue angle. Even if the center value of the hue angle is not an integer but a decimal number, the center value of the hue angle may be adopted as a representative value of the hue angle. When the central value of the hue angle is a decimal, the amount of calculation can be reduced by adopting an integer near the central value of the hue angle as a representative value of the hue angle.

  When it is necessary to further reduce the amount of calculation, a simple number may be assigned as a representative value of each hue angle for convenience. That is, in the case of this example, the hue angle of the representative value “30” can be expressed as a representative value “1”, and subsequently expressed as “2”, “3”.

  As described above, in the color space region dividing unit 91c according to the first embodiment, a region composed of continuous hue angle values is set as one class. Therefore, each frequency of the hue histogram shown in FIG. This is the sum of signals of hue angles up to degrees. For example, the frequency of the hue histogram Dbh (30) indicated by the numerical value 30 on the horizontal axis corresponds to the sum of signals from the hue angle 0 degree to the hue angle 60 degrees included in the luminance signal Db for one frame. Hereinafter, the hue of the hue angle with the representative value “i” is referred to as Cbh (i), and the frequency is referred to as Dbh (i). That is, in the example of FIG. 3, since the sum of signals of hue angles from 0 to 60 degrees is represented as a representative value “30”, the hue of the region is represented as Cbh (30) and the frequency is represented as Dbh (30).

  Accordingly, the color space region dividing unit 91c performs the hue histogram Dbh (30), the hue histogram Dbh (90), the hue histogram Dbh (150), the hue histogram Dbh (210), the hue histogram Dbh (270), and the hue histogram Dbh (330). Are divided into six types of hue histograms Dbh. That is, the color space area dividing unit 91c obtains hue histograms Dbh (30) to Dbh (330) in six divided areas based on the hue information H obtained from the image signal Db.

  In the example shown in FIG. 3, the hue information H is used to create the hue histogram Dbh. However, even if the histogram is created using the saturation information S and the brightness information V other than the hue information H, It doesn't matter. Further, it is possible to divide into a plurality of divided areas in the color space by combining at least two of the hue information H, the saturation information S, and the lightness information V, and create histograms in the divided areas.

  In Embodiment 1, the hue information H is divided into a plurality of regions by creating a histogram related to the hue information H, but the region is divided by a method other than creating a histogram. It is also possible.

  In addition, regarding the division of the region, repeated division may be performed. For example, after creating a histogram related to the hue information H, depending on the result, for the hue information H, the saturation information S, and the lightness information V, the region is divided again to perform more detailed analysis. Can do.

  Unlike the hue histogram of FIG. 3, the hue histogram may be generated by counting the angle for each hue angle. That is, each hue angle may be constituted by one hue angle. In this case, the representative value of each hue angle is the value of the hue angle that constitutes the hue angle.

  Further, when the hue angle is divided, the number of divisions may be set to other than 6, and the division number and range of the hue histogram may be set freely. For example, the hue angles “0” to “60” and “300” to “360” can be set in increments of 10 degrees, and the others can be set in increments of 30 degrees. Alternatively, the frequency of unnecessary hue angles may be omitted. In this way, by changing the number of divisions and omitting the frequency of unnecessary hue angles, the calculation amount in the color space region division unit 91c and the accuracy of the hue histogram can be adjusted. The number of divisions and the range may be selected according to the characteristics of the content to be detected and the content genre.

  In this example, the frequency of the hue histogram is represented by the number of pixels with respect to the hue angle of the image signal Db for one frame, but may be displayed as the cumulative number of pixels. Further, 10% or the like may be displayed as a ratio of the total number of pixels, not the number of pixels.

  FIG. 4 is an explanatory diagram illustrating an example of generation of the hue signal feature value Ch by the hue information analysis unit 92c. As shown in the figure, the hue information analysis unit 92c has a threshold value in which each frequency Dbh (i) of the hue histogram Dbh is smaller than a predetermined hue determination threshold value TDbh2, or a threshold value TDbh2 and a predetermined threshold value larger than the threshold value TDbh2. It is determined whether the value is equal to or less than TDbh1 (> TDbh2) or greater than the hue determination threshold value TDbh1, and one of the three classification information values, small (low), medium, and large (high) is selected. A prescribed hue determination information value Dbhc (i) is calculated.

  In the example shown in FIG. 4, the hue histogram Dbh (30) is a value larger than the threshold value TDbh1, and therefore, a value indicating “large (high)” is input as the hue determination information value Dbhc (30). Further, since the hue histogram Dbh (90) is a value between the threshold value TDbh2 and the threshold value TDbh1, a value indicating “medium” is input as the hue determination information value Dbhc (90). Since the hue histograms Dbh (150), Dbh (210), Dbh (270), and Dbh (330) are smaller than the hue determination threshold TDbh2, the hue determination information values Dbhc (150), Dbhc (210), Dbhc (270) ) And Dbhc (330) are input with values indicating "small (low)".

  In this example, the threshold value for frequency Dbh (i) comparison is two types (TDbh2, TDbh1), which are classified into three types of large, medium, and small. That is, the threshold value may be more or less than two types. By adjusting a large number of threshold values, it is possible to classify the hue determination information value Dbhc in more detail. Further, the processing amount can be reduced by reducing the number of thresholds.

  Further, the threshold value may be changed for each hue histogram Dbh (i). For example, the hue histogram Dbh (30) is determined based on only the predetermined threshold value TDbh1, and the hue histogram Dbh (90) obtains the hue signal feature value Ch based on the predetermined threshold value TDbh1 and other predetermined threshold values TDbh2 and TDbh3. You can also take the method. In this way, by changing the threshold value for each hue histogram Dbh (i), it is possible to focus on the hue to be analyzed in detail, and to reduce the analysis accuracy of hues that do not require detailed analysis. Obtainable.

  The hue information analysis unit 92c uses the hue signal feature value Ch defining the hue determination information value Dbhc (i) in the hue angle region Cbh (i) (hue histogram Dbh (i)) as the hue information value Ci, and the content feature determination unit 10 is output.

  That is, in the example shown in FIG. 4, Dbhc (30) = “large”, Dbhc (90) = “medium”, Dbhc (150) = Dbhc (210) = Dbhc (270) = Dbhc (330) = “small”. The specified hue signal feature value Ch is output to the content feature determination unit 10 as the hue information value Ci.

  Therefore, the color space region dividing unit 91c performs statistical classification processing on six divided regions (hue histogram Dbh (30), etc.) on the hue information including the hue information H to obtain a hue signal feature value Ch that is a statistical result. Thus, it is possible to obtain a hue signal feature value Ch with high accuracy by analyzing the hue information H in detail.

  In outputting the hue signal feature value Ch, by specifying one of the hue angle region Cbh (i) and the hue determination information value Dbhc (i), only the hue angle region Cbh (i) or The hue signal feature value Ch that defines only the hue determination information value Dbhc (i) may be output.

  For example, a method of outputting a hue signal feature value Ch that defines a hue angle region Cbh (i) where the hue determination information value Dbhc (i) = “large” is conceivable. Returning to the example of FIG. 4, the hue signal feature value Ch defining the hue angle region Cbh (30), which is the hue angle region Cbh (i) satisfying Dbhc (i) = “large”, is the hue information value Ci. Is output to the content feature determination unit 10.

  As another example, a method of outputting a hue signal feature value Ch that defines the hue determination information value Dbhc (30) in the hue angle region Cbh (30) is conceivable.

  Returning to the example of FIG. 4, the content feature determination unit 10 uses the hue signal feature value Ch that defines the hue determination information value Dbhc (30) = “large” corresponding to the hue angle region Cbh (30) as the hue information value Ci. Is output.

  FIG. 5 is an example of a block diagram illustrating a detailed configuration of the content feature determination unit 10. As shown in the figure, the content feature determination unit 10 in the image display apparatus according to the first embodiment includes a hue feature determination unit 101, a multi-frame integration unit 102, and a determination speed adjustment unit 103.

  The hue feature determination unit 101 is based on the combination content defined by the hue information value Ci (hue signal feature value Ch) output from the hue information analysis unit 9, and the content feature determination value Si in the image signal Db for one frame. (Temporary content feature determination value) is determined and output to the multi-frame integration unit 102 and the determination speed adjustment unit 103, respectively.

  FIG. 6 is an explanatory diagram showing the determination content of the content feature determination value Si by the hue feature determination unit 101 in a table format. In the example shown in the figure, six types of values Dbhc (30), Dbhc (90), Dbhc (150), Dbhc (210), Dbhc (270) defined by the hue information value Ci (hue signal feature value Ch) are shown. ) And Dbhc (330) values (large, medium, or small) are combined in 729 ways (the sixth power of 3), and 729 types of content feature judgment values Si including feature judgment values S1 to S729 are used as hue features. The determination unit 101 determines.

  It should be noted that the content feature determination combination by the hue feature determination unit 101 can be arbitrarily created based on the viewer's preference and the video database. In the example of FIG. 6, the hue determination information values Dbhc (30), Dbhc (90), Dbhc (150), Dbhc (210), and Dbhc (270) are “small”, and the hue determination information value Dbhc (330). Is “large”, the hue feature determination unit 101 selects the feature determination value S3 as the determination type. Then, the content feature determination value Si instructing the feature determination value S3 is output from the hue feature determination unit 101 to the multiple frame integration unit 102 and the determination speed adjustment unit 103, respectively.

  Further, the hue feature determination unit 101 may make a determination based on some information of the six types of combination contents defined by the hue information value Ci (hue signal feature value Ch). For example, the hue feature is determined based only on the hue determination information value Dbhc (30). It is possible to select determination contents such as determining hue characteristics based on a combination of hue determination information values Dbhc (30) and Dbhc (90). Thus, by reducing the amount of information used for the hue feature determination processing by the hue feature determination unit 101 among the information defined by the hue information value Ci (hue signal feature value Ch), the hue feature determination speed is increased, The required memory capacity can be reduced.

  The content feature determination unit 10 shown in FIG. 5 is a configuration example, and the content feature determination unit 10 can determine the content feature information value Si based on the color information value Ci (hue signal feature value Ch). Anything is acceptable.

  As an example of a method of outputting the content feature information value Si based on the color information value Ci, a method of calculating the likelihood of each color information value Ci and outputting the content feature information value Si from statistical processing, etc. Can be considered.

  Further, as an example of a method for obtaining the content feature information value Si based on the hue information value Ci, a histogram relating to a hue signal obtained from an image signal for one frame is created, and the hue information value Ci configured from the corresponding histogram is generated. On the basis of this, a method of calculating a new hue information value Ci by obtaining an average value for a plurality of frames and determining the feature of the video content with respect to the image signal can be considered. In this manner, the hue information value Ci obtained from the image signal for one frame may be used to obtain a new hue information value Ci by calculation for a plurality of frames, and the feature determination of the video content with respect to the image signal may be performed.

  FIG. 7 is an explanatory diagram showing other determination contents of the content feature determination value Si by the hue feature determination unit 101 in a table format. FIG. 7 shows an example in which the content feature determination value Si obtained by the hue feature determination unit 101 is classified into six types (S1 to S5, S0) for each video genre in all cases.

  In this example, the hue information value Ci is reduced from the above example, and the content feature determination value Si is determined based on the hue determination information values Dbhc (30), Dbhc (90), Dbhc (150), and Dbhc (210). The content feature determination value Si is classified into six types for each video genre.

  The classification shown in this example is a database created by examining video contents of various genres, statistically classifying distribution features, and associating with each content feature information value Si. In this example, the genres are classified into six types, but any number of classifications may be used. In the figure, the feature determination values S0 to S5 are output as the feature information value Si of the content by the hue feature determination unit 101. For example, if the hue determination information value Dbhc (30) is “small”, Dbhc (90) is “medium”, Dbhc (150) is “medium”, and Dbhc (210) is “small L”, the content feature determination value Si Is determined to be S1 (sports) and output.

  In the above example, since classification is performed by six video genres, the classification name is referred to as a representative genre name. In the classification of FIG. 7, sports (S1) is mainly a genre of sports and natural images, and has a feature that the hue distribution is green and blue in a video such as soccer or baseball green. The music (S2) is mainly a genre such as a music clip of music, and there are many images using light and dark and many vivid colors. Studio (S3) is mainly a variety and media genre, and there are many videos shot in the studio. For this reason, it has a feature that the average luminance is high and there are many bright colors as a whole. The movie (S4) is mainly characterized in that the genre is a movie and there are many dark scenes. The drama (S5) is mainly a genre of drama and animation, and has the characteristics that there are few blacks but low average brightness and many dark shades. Normal (S0) is a video with few features that makes it difficult to determine the genre.

  Since the genre (content feature determination value Si) determined by the hue feature determination unit 101 in this example is determined based on the image signal Da that defines the video content, each of the content feature determination values Si is representative. It shows that it is close to the video feature of the genre. For example, if the input video is a movie with the theme of sports, it is classified as sports (S1), not movies (S4).

  In content feature detection, in order to improve the accuracy of genre discrimination, it may be used in combination with other known methods. Examples of other methods include genre information of a digital program guide as disclosed in Patent Document 1.

  As described above, the hue feature determination unit 101 statistically uses the hue information H defined by the hue information value Ci (hue signal feature value Ch) obtained by the hue information analysis unit 9 based on the image signal Db. By classifying the features and determining the content feature determination value Si, it is possible to accurately determine the content feature and genre of the video. The information specified by the hue information value Ci is information related to at least one of the hue information H, the saturation information S, and the lightness information V (however, only the lightness information V is excluded). I can expect.

  The multi-frame integration unit 102 performs an analysis process based on the input content feature information value Si, and based on the determination speed information value Ti output from the determination speed adjustment unit 103, a plurality of frame content feature information values Si. And a plurality of content feature information values Fi reflecting a plurality of content feature values Si are obtained. This makes it possible to perform content determination with higher accuracy and more stable as compared to the content feature information values Si of a plurality of frames.

  Further, the determination speed adjustment unit 103 outputs the determination speed information value Ti to the multiple frame integration unit 102, and adjusts the determination speed by the multiple frame integration unit 102, thereby determining the content according to the feature of the content. To adjust the timing for gradation correction.

  An example of an analysis processing method in the multi-frame integration unit 102 will be described. The multi-frame integrating unit 102 adds the feature determination values Si (determination type) for one frame based on the content feature determination values Si sequentially input, and the same feature determination value (determination type) is determined as the determination speed adjustment unit 103. There is a technique (first determination example) for determining as the multiple feature determination value Fi when the predetermined determination number N1 (predetermined number of appearances) indicated by the determination speed information value Ti output by the above is reached.

  Specifically, for example, when the fixed determination number N1 indicated by the determination speed information value Ti is “10”, the content feature determination value Si for each frame is added for a plurality of frames as shown in FIG. In FIG. 8, the content feature determination value that has reached the fixed determination number N1 = 10 earliest is “S3”. Therefore, the multiple content feature determination value Fi is detected as “S3”, and the feature determination value S3 is The instructed multiple content feature information value Fi is input to the determination speed adjustment unit 103 and the correction control unit 4 in the image correction unit 6.

  As another example of the technique, the multi-frame integration unit 102 adds the feature determination values Si for one frame based on the sequentially input content feature determination values Si, and the same feature determination values continue. Thus, when the constant determination number N2 (predetermined number of consecutive appearances) indicated by the determination speed information value Ti output from the determination speed adjustment unit 103 is reached, a method (second determination) is determined as the multiple feature determination value Fi. Example).

  Specifically, for example, when the fixed determination number N2 is “10”, the content feature determination value Si for each frame is added for a plurality of frames as shown in FIG. In this example, the content feature determination value is “S3” 10 times continuously from the 4th frame to the 13th frame. For this reason, the content feature determination value that has reached the predetermined determination number of 10 consecutively earliest is “S3”. As a result, the multiple content feature determination value Fi is determined to be S3 by the multiple frame integration unit 102, and the multiple content feature information value Fi indicating the feature determination value S3 is determined by the determination speed adjustment unit 103 and the image correction unit 6. Is input to the correction control unit 4.

  As another method, the multi-frame integration unit 102 adds the feature determination value Si for one frame to the input content feature determination value Si, and the same feature determination value is determined as the determination speed adjustment unit 103. A method (third determination example) for determining as the multiple feature determination value Fi when the maximum appearance degree is within the predetermined determination number N3 (predetermined period) indicated by the determination speed information value Ti output by .

  Specifically, for example, when the fixed determination number N3 is “15”, the content feature determination value Si for each frame is added for 15 frames, which is the fixed determination number N3, as shown in FIG. In this example, the appearance level of each content feature determination value in 15 frames is “S1” 0 times, “S2” 7 times, “S3” 6 times, “S4”, “S5” 1 each. Times. The feature determination value having the maximum appearance degree is the feature determination value type “S2” for which determination has been performed seven times. Therefore, a plurality of content feature determination values Fi instructing the feature determination value S 2 are input to the determination speed adjustment unit 103 and the correction control unit 4 in the image correction unit 6 by the multiple frame integration unit 102.

  As described above, the multi-frame integration unit 102 calculates the determination types indicated by the sequentially obtained content feature determination values Si in accordance with the first to third determination examples described above, and calculates the content feature determination values Si for a plurality of frames. By analyzing a considerable amount, it is possible to obtain a plurality of content feature information values Fi that are more accurate than the content feature determination value Si.

  In addition, the multi-frame integrating unit 102 appropriately determines the timing for obtaining the multi-content feature determination value Fi by determining the values of the predetermined determination numbers N1 to N3 that are equivalent to predetermined frames based on the determination speed information value Ti. Can be adjusted. As described above, the determination speed information value Ti is determined by the determination speed adjustment unit 3 based on the content feature determination value Ji and the multiple content feature determination value Fi.

  As a result, the multi-frame integration unit 102 determines the determination speed (determination based on the fixed determination numbers N1 to N3, etc.) that matches the feature of the video content such as the genre indicated by the content feature determination value Si and the multiple content feature determination value Fi. A multiple content feature determination value Fi can be obtained.

  Further, the plurality of frame integration units 102 may be implemented in combination of a plurality of analysis processes represented by the above three methods (first to third determination examples).

  For example, the feature judgment value Si for one frame is added as the constant judgment number Na, and after the same feature judgment value reaches the constant judgment number Na, the feature judgment value Si for one frame is added as the constant judgment number Nb. For example, when the same feature determination value continuously reaches a certain determination number Nb, a combination of determining as a plurality of content feature determination values Fi for the first time is conceivable.

  As another combination method, the feature determination value Si for one frame is added as the fixed determination number Na to obtain the multiple content feature determination value Fai, and then the multiple content feature determination value Fai is added to obtain the multiple content feature. The determination value Fi may be obtained. In this example, the combination is two times, but two or more combinations may be performed.

  As described above, the multi-frame integrating unit 102 obtains the multi-content feature information value Fi using a plurality of combinations including at least one determination criterion in the first to third determination criteria described above based on the content feature determination value Si. Thus, when the multiple content feature determination value Fi is determined, it is possible to adjust the speed and accuracy until the determination is made. As a result, it is finally possible to perform gradation correction with high adaptability by the image correction unit 6.

  The multiple frame integration unit 102 according to the first embodiment may output the content feature determination value Si as the multiple content feature determination value Fi. In other words, the processing by the multiple frame integration unit 102 is omitted, and the content feature determination value Si output from the hue feature determination unit 101 is directly taken into the determination speed adjustment unit 103 and the correction control unit 4 as the multiple content feature determination value Fi. Can also be adopted.

  In the multi-frame integration unit 102, when the multi-content feature information value Fi is not fixed such as immediately after the power is turned on, a genre such as normal is used as a genre indicating that the multi-content feature information value Fi is indefinite. You may set your own. In this example, “normal” is set as a genre having no distinctive features.

  Even when the content of only one frame within the same genre changes significantly in the same genre, the determination type of the content feature determination value Si in the frame is the multiple frame integration unit 102 according to the first embodiment. It is automatically removed by the analysis processing in 102, and it is possible to prevent extreme image quality correction from being performed by the image correction unit 6 based on the multiple content feature determination value Fi.

  That is, in the case of a video in which the video content is switched for each frame, the multi-frame integration unit 102 can prevent the image quality from being switched for each frame, thereby preventing an unnatural video.

  Based on the content feature information value Si output from the hue feature determination unit 101 and the multiple content feature information value Fi output from the multiple frame integration unit 102, the determination speed adjustment unit 103 is optimal for the changing process of content and genre. The determination speed information value Ti for instructing the constant determination number N1 or the like that is considered to be calculated is output to the multi-frame integration unit 102.

  FIG. 11 is an explanatory diagram showing an example of creation of the judgment speed information value Ti by the judgment speed adjustment unit 103 in a table format. As shown in the figure, a determination speed information value Ti corresponding to each combination of the content feature information value Si and the plurality of content feature information values Fi is obtained.

  Specifically, for example, when the content feature information value Si is a feature determination value S3 and the multiple content feature information value Fi is a value indicating the multiple feature determination value F4, the determination speed information indicating the determination speed information value T43. The value Ti is output to the multi-frame integration unit 102.

  For example, when the hue feature determination unit 101 outputs the content feature determination value Si in the combination shown in FIG. 7, the combination of the multiple feature determination value F3 and the feature determination value S3 is like a movie indicated by the multiple feature determination value (F4). The timing of changing from a genre having a large overall dark hue to a genre having a high average luminance and a bright scene such as variety and news indicated by the feature determination value (S3) is indicated.

  In this case, since the image correction performed in the image correction unit 6 is performed based on the multiple feature determination value (F4) that is the current determination content, the features of the image correction are opposite to each other. For this reason, the determination speed information value T43 is set so as to indicate a relatively high speed value, thereby setting an unnatural image correction period to be as short as possible so that the unnatural image correction does not stand out. be able to.

  In the above example, the determination speed adjustment unit 103 obtains the determination speed information value Ti based on the content feature information value Si and the multiple content feature information value Fi, but only the multiple content feature information value Fi or the content feature information. The determination speed information value Ti may be adjusted using only the value Si.

  The determination speed adjustment unit 103 may have any configuration as long as the determination speed information value Ti can be adjusted based on the characteristics included in the image signal Da. Further, the determination speed information value Ti may be adjusted by a signal other than an image signal such as an audio signal.

  Returning to FIG. 1, the correction control unit 4 selects a correction parameter Pi corresponding to the multiple content feature information value Fi and outputs it to the correction execution unit 5. For example, when the content feature information value Si indicates the feature determination value S1 by a combination of the tint information value Ci, and finally the multiple content feature information value Fi indicates the multiple feature determination value F1, the multiple feature determination value F1 is supported. The correction parameter P1 to be selected is selected. The correction parameter Pi can be freely adjusted according to the viewer's preference and system based on the trend and characteristics of the genre.

  The correction parameters Pi for the multiple content determination values Fi output from the correction control unit 4 to the correction execution unit 5 are, for example, luminance control based on video contrast, sharpness control based on sharpness, color density, and three-dimensional (3D) noise. These include noise removal control by reduction, luminance correction by gamma correction, and backlight brightness adjustment.

  FIG. 12 is an example of the output of the correction control unit 4 for the multiple content feature determination values Fi classified into five genres. The setting of the correction parameter Pi of the correction control unit 4 is, for example, when the value indicated by the multiple content feature determination value Fi is sport (F1), the luminance is mostly halftone as the video feature defined by the image signal Db. It can be mentioned that the gradation is spread to all gradations on average, and there are many images in which the hue is mostly green and blue. Therefore, the gamma correction is linear with the x-axis as the gradation value of the input signal (image signal Db) and the y-axis as the output gradation value (image signal Dc), and the contrast is “medium”. Because of the natural image, sharpness is “low”, color depth is “medium”, and backlight brightness is “medium”. Since it is a sport, 3D noise reduction of 3D processing can be tailored in a moving image, so that it can be set to “low” to cope with intense movement.

  When the value indicated by the multiple content feature determination value Fi is, for example, a movie (F4), where there are many dark scenes as a whole and the influence of reducing the brightness of the backlight is small, the backlight Energy saving of the image display apparatus can be achieved by setting the brightness to “low”. Thus, by reducing the brightness of the backlight in accordance with the content of the video content, it is possible to save energy while maintaining a certain level of image quality.

  Here, the correction parameters Pi to be prepared may be provided by the number of combinations based on the color information value Ci, or may be the same as the number of types of content features.

  The correction execution unit 5 performs image correction on the image signal Db based on the correction parameter Pi. This image correction is performed for each frame.

  The correction execution unit 5 may start image correction from the video signal Db analyzed when the multiple content feature determination value Fi is obtained by delaying the video signal Db by using a frame buffer or the like. Alternatively, image correction may be started from a newly obtained frame after the image signal Db analyzed when the multiple content feature determination value Fi is obtained.

  Further, the correction execution unit 5 may perform image correction in accordance with the scene change. Similarly, the parameter amount to be corrected may be set to a very small value, and the correction may be performed in several times during the fixed time M. In this case, it is conceivable that the value of the predetermined time M is also adjusted according to the multiple content feature determination value Fi. By adjusting the correction timing in this way, it is possible to perform natural image correction that does not give the viewer a sense of incongruity.

  As described above, the content feature determination unit 10 in the image display apparatus according to the first embodiment obtains the multiple content feature determination value Fi that is the determined content feature determination value based on the color information value Ci. The tint information value Ci is obtained from the tint information analysis unit 9 by performing analysis processing on a plurality of divided areas divided in the color space with respect to the tint information including the hue information.

  Therefore, the image display apparatus according to the first embodiment uses the image correction unit 6 to generate the image signal Db for one frame based on the multiple content feature determination value Fi that accurately reflects the characteristics of the content obtained from the image signal Db. By performing the image correction, it is possible to perform the image correction with the deterioration of the image quality suppressed by adapting to the feature of the content or the change process of the feature while maintaining the real-time property. This effect can be expected if the information defined by the hue information value Ci is information related to at least one of the hue information H, the saturation information S, and the lightness information V (however, only the lightness information V is excluded). .

  As described above, in the image display device according to the first embodiment, based on the content feature determination value Si for one frame determined from the hue information value Ci including the hue information H, the feature amount for a plurality of frames is used. Since the characteristics of the content are determined and the multiple content feature determination value Fi is obtained, the characteristics of the content can be determined more accurately.

  That is, the multi-frame integration unit 102 in the content feature determination unit 10 in the image display apparatus according to the first embodiment outputs the hue information value Ci, which is a temporary content feature determination value obtained from the hue feature determination unit 101, for a predetermined frame. Since the plurality of content feature determination values Fi are obtained by performing a considerable amount analysis, it is possible to accurately determine the characteristics of the content obtained from the image signal Db.

  Further, since the multi-frame integrating unit 102 changes the content determination speed based on the determination speed information value Ti, the timing for determining the content is adjusted according to the feature of the content, and the image correction by the image correction unit 6 is performed. Can be adjusted. As a result, it is possible to perform image correction in which a change in image quality is not noticeable.

  The image display device according to the first embodiment is not limited to the video display device and can be used in other fields related to video as a method for more accurately determining the feature amount of the video content from the color information value Ci. For example, a video recording device such as a video recorder such as a hard disk or a DVD may be used as another field.

  On the other hand, in order to perform the content feature determination using the technique described in Patent Document 2 described above, it is necessary to hold multimedia data whose content type is known in advance. . For this reason, classification cannot be performed for first-time programs.

  In Patent Document 3, since the video signals are classified only by the luminance histogram based on the lightness signal V, the video signals are classified into a plurality of divided regions based on the hue information H or the saturation information S included in the image signal Db. Compared with, classification accuracy is inferior, and detailed and accurate classification is not possible.

  In the image display device according to the first embodiment, the color space is divided into a plurality of color spaces based on the HSV color space. For example, the RGB color space, the CMY color space, the HLS color space, the YCbCr color space, the Lab color system, The present invention can also be applied to all other color spaces and color systems such as the L * a * b * color system, the L * u * v * color system, and the like.

<Embodiment 2>
FIG. 13 is a block diagram showing a configuration of an image display apparatus according to Embodiment 2 of the present invention. The image display device according to the second embodiment includes the image processing device 17 in place of the image processing device 7 in the image processing device according to the first embodiment described above.

  The image processing apparatus 17 according to the second embodiment includes a color information analysis unit 19, a content feature determination unit 12, and an image correction unit 6. The image correction unit 6 includes a correction control unit 4 and a correction execution unit 5. Have. The image signal Db obtained from the reception unit 2 is input to the color information analysis unit 19 and the correction execution unit 5 of the image processing device 17, respectively.

  The hue information analysis unit 19 detects the hue information value Ci in each frame from the hue information H, the saturation information S, and the brightness information V included in the image signal Db, and outputs them to the content feature determination unit 12.

  The content feature determination unit 12 determines the feature of the video content for one frame based on the color information value Ci, analyzes the video content for a plurality of frames, and then, as a result of the analysis, the multiple content feature determination value Fi is output to the correction control unit 4.

  The correction control unit 4 calculates a correction parameter Pi used when the correction execution unit 5 performs image correction on the image signal Db for one frame based on the multiple content feature information Fi, and outputs the correction parameter Pi to the correction execution unit 5.

  FIG. 14 is a block diagram illustrating details of an internal configuration example of the color information analysis unit 19 illustrated in FIG. 13. As shown in FIG. 14, the color information analyzing unit 19 in the image display apparatus according to the second embodiment includes a color information detecting unit 190c, a color space region dividing unit 191c, and a color information analyzing unit 192c. Yes.

  In the example illustrated in FIG. 15, the image signal Db obtained from the receiving unit 2 and including the hue information H, the saturation information S, and the lightness information V is input to the hue information detection unit 190c.

  The hue information detection unit 190c extracts the hue information H, the saturation information S, and the lightness information V from the image signal Db for one frame, and outputs them to the color space region division unit 191c.

  The color space region dividing unit 191c divides the color space into a plurality of color regions based on the hue information H, the saturation information S, and the lightness information V in the image signal Db for one frame, and the region is divided from the color regions. Information JDd (H, S, V, D) is detected and output to the color information analysis unit 192c. In the second embodiment, the HSV color space is used as an example of the color space.

  The color information analysis unit 192c calculates a color characteristic value Ct in the image signal Db for one frame based on the area division information JDd generated by the color space area division unit 191c, and sets it as the color information value Ci. The color information value Ci is output to the content feature determination unit 12.

  In the example shown in FIG. 14, when the input video signal Db is an interlace signal, the color information detection unit 190c converts the image data for two fields into an image signal for one frame, and obtains the area division information JDd. As described above, the color information detection unit 190c may perform preprocessing for obtaining the region division information JDd.

  As another example of pre-processing, when the image signal Db is shown in a color system different from the hue information H, the saturation information S, and the lightness information V such as an RGB signal, the hue information H is expressed by a known calculation formula. It is also conceivable to convert into saturation information S and lightness information V.

  FIG. 15 shows an example of area division by the color space area dividing unit 191c. In the example shown in FIG. 15, the region division information JDd is created by dividing the HSV color space by the hue information H, the saturation information S, and the lightness information V. In this example, in dividing the HSV color space, after dividing by the saturation information S (first division), the hue information H and the lightness information V are divided in detail (second division).

  In dividing the HSV color space using the saturation information S, the color space region dividing unit 191c determines whether the saturation information value S is smaller than a predetermined saturation determination threshold value r1 or the threshold value r1. It is determined whether it is a value between the threshold value r2 which is a large predetermined threshold value or a value larger than the saturation threshold value r2, and each of the three divided regions Dr (1), Dr (2), Dr (3) is determined. The first division is performed.

  In the example shown in FIG. 15, the saturation information S is divided into three types based on threshold values, but the number of divisions may be at least as large. Similarly, the number of divisions based on the lightness information V may be larger or smaller than the example shown in FIG. Further, in the division, if any one or more of the hue information H, the saturation information S, and the lightness information V are used, the other information may be omitted. This reduces the accuracy of classification, but enables high-speed computation. However, it is desirable to include at least one of the hue information H and the saturation information S.

  FIG. 16 is an explanatory diagram showing a first example of the second division. As shown in the figure, the divided region Dr (1) after the first division is further divided by the brightness information V. As shown in FIG. 16, in the first example of the second division, the division information Dr (1) is divided into six equal parts so that the lightness information V is uniform, and six types of division areas Dd (1) to Dd are further divided. The division to obtain (6) is performed. The divided region Dr (1) is a near-monochrome image region with no vivid hue because the saturation information is smaller than the threshold value r1. Therefore, the divided region Dd (1) is close to white, and the color changes to black as the divided region Dd (6) approaches.

  In the example shown in FIG. 16, the division is performed so that the lightness information V is uniform. However, the division may be performed non-uniformly, and the number of divisions is not limited to six, and may be freely adjusted. PCCS (Practical Color Coordinate System) can be used as a reference for dividing. Further, since the saturation information S is small in the divided area Dr (1), the influence of the hue information H is reduced. Therefore, in this example, the division by the hue information H is not performed, but the division information Dr (1) may be further divided in more detail using the hue information H.

  FIG. 17 is an explanatory diagram showing a second example of the second division. As shown in the figure, the divided region Dr (2) after the first division is further divided by the hue information H. As shown in FIG. 17, in the second example of the second division, the hue information H is divided into six equal parts from the divided region Dr (2), and six types of divided regions Dd (7) to Dd are further divided. The division to obtain (12) is performed. The divided area Dr (7) represents a hue angle of 0 to 60 degrees and includes a hue that changes from red to yellow, and the divided area Dr (8) represents a hue angle of 60 to 120 degrees and changes from yellow to green. The divided region Dr (9) represents a hue angle of 120 degrees to 180 degrees, and includes a hue that changes from green to light blue. Further, the divided region Dr (10) represents a hue angle of 180 degrees to 240 degrees and includes a hue that changes from light blue to blue, and the divided region Dr (11) represents a hue angle of 240 degrees to 300 degrees, from blue A hue that changes from purple is included, and the divided region Dr (12) represents a hue angle of 300 degrees to 360 degrees, and includes a hue that changes from purple to red.

  In the example shown in FIG. 17, the division is performed so that the hue information H is uniform, but the division may be performed non-uniformly, and the number of divisions is not limited to six, and may be freely adjusted. Further, the divided area Dr (2) may be further divided in detail using the lightness information V.

  FIG. 18 is an explanatory diagram showing a third example of the second division. As shown in the drawing, the divided region Dr (3) after the first division is further divided by the hue information H. As shown in FIG. 18, in the third example of the second division, the hue information H is divided into six equal parts from the divided region Dr (3), and six types of divided regions Dd (13) to Dd are further divided. The division to obtain (17) is performed. The divided area Dd (13) represents a hue angle of 0 to 60 degrees and includes a hue that changes from red to yellow, and the divided area Dd (14) represents a hue angle of 60 to 120 degrees and changes from yellow to green. The divided region Dd (15) represents a hue angle of 120 degrees to 180 degrees, and includes a hue that changes from green to light blue. Further, the divided area Dd (16) represents a hue angle from 180 degrees to 240 degrees, and includes a hue that changes from light blue to blue. The divided area Dd (17) represents a hue angle from 240 degrees to 300 degrees, and from blue A hue that changes from purple is included, and the divided region Dd (18) represents a hue angle of 300 degrees to 360 degrees, and includes a hue that changes from purple to red.

  In the example illustrated in FIG. 18, the division is performed so that the hue information H is uniform, but the division may be performed non-uniformly, and the number of divisions is not limited to six and may be freely adjusted. Further, the divided area Dr (3) may be further divided in detail using the lightness information V.

  Regarding the division of the region, it may be repeated. For example, as shown in FIGS. 15 to 18 described above, after dividing the HSV color space into 18 types (detailed divided regions Dd (1) to Dd (18)), some regions are further divided depending on the result. By doing so, detailed analysis can be performed. Note that some regions include two or more regions that have been divided.

  The color space region dividing unit 191c calculates the frequency of each divided region Dd (i), that is, the number of pixels for the divided region of the image signal Db for one frame, and the hue information H, saturation information S, Along with the lightness information V, information including the frequency is output as region division information JDd (H, S, V, D) to the color information analysis unit 192c.

  In this example, the frequency of the divided region Dd (i) is represented by the number of pixels with respect to the divided region of the image signal Db for one frame, but may be displayed as the cumulative number of pixels. Further, 10% or the like may be displayed as a ratio of the total number of pixels, not the number of pixels.

  Further, when the region is divided, the number and range of division may be freely set. For example, in the case of division based on the hue information H, the hue angles “0” to “60” and “300” to “360” may be in increments of 10 degrees, and the others may be in increments of 30 degrees. Alternatively, the frequency of unnecessary hue angles may be omitted. In this way, by changing the number of divisions and omitting the frequency of unnecessary hue angles, the calculation amount by the color space region division unit 191c and the accuracy of the division region Dd (i) can be adjusted. The number of divisions and the range may be selected according to the characteristics of the content to be detected and the content genre.

  The color information analysis unit 192c determines whether each frequency Ddh (i) of the divided area Dd (i) defined by the area division information JDd is a value smaller than a predetermined color information determination threshold value TDCh2 or a threshold value TDCh2 It is determined whether the value is a value that is larger than or equal to the color information determination threshold value TDCh1 that is a predetermined threshold value or greater than the color information information determination threshold value TDCh1. Then, the hue information analysis unit 192c indicates the color information determination information value DDCc (i) indicating any one of the three classification information values, “small (low)”, “medium”, and “large (high)”. ) Is calculated. The color information analyzing unit 192c outputs the color characteristic value Ct that defines the instruction content of each color information determination information value DDCc (i) as the color information value Ci.

  In this example, there are two predetermined thresholds, and the color information determination information value DDCc (i) is classified into three types of large, medium, and small, but may be classified into other than three types. That is, the threshold value may be more or less than two types. More detailed classification is possible by adjusting a large number of threshold values. Further, the processing amount can be reduced by reducing the number of thresholds.

  Further, the threshold value may be changed for each divided region Dd (i). For example, the detailed divided region Dd (1) (Ddh (1)) is determined based on only a predetermined threshold value TDCh1, and Ddh (2) is colored based on a predetermined threshold value TDCh1 and other predetermined threshold values TDCh2 and TDCh3. A technique of obtaining the feature value Ct can also be taken. As described above, by changing the threshold value according to the divided region Dd (i), the HSV color space to be analyzed in detail is focused on, and the analysis accuracy of the space that does not require detailed analysis is reduced. Can be obtained.

  In this way, the hue information analysis unit 192c provides the content feature determination unit 12 with the color feature value Ct that defines the color information determination information value DDCc (i) in each divided region Dd (i) as the color information value Ci. Output.

  In other words, the color information analysis unit 192c analyzes results in the detailed divided regions Dd (1) to Dd (18) in which the color space is divided by at least two combinations of the lightness information V, the hue information H, and the saturation information S. Therefore, the color feature value Ct that more reflects the content feature of the image signal Db can be obtained.

  As a result, the image display device according to the second embodiment uses the image correction unit 6 based on the multiple content feature determination value Fi output from the content feature determination unit 12 based on the color information value Ci that is the color feature value Ct. Since image correction is performed on the image signal Db, there is an effect that image correction with higher accuracy can be performed.

  Note that when outputting the color feature value Ct, one of the divided region Dd (i) and the color information determination information value DDCc (i) may be defined. That is, the hue information analysis unit 192c may output the color feature value Ct that defines only the divided region Dd (i) or only the color information determination information value DDCc (i).

  For example, a method of outputting the color feature value Ct that defines the divided area Dd (i) where the color information determination information value DDCc (i) = “large (high)” is conceivable. As another example, a method of outputting the color information determination information value DDCc (1) of the divided region Dd (1) as the color characteristic value Ct is conceivable.

  FIG. 19 is an example of a block diagram illustrating a detailed configuration of the content feature determination unit 12. As shown in the figure, the content feature determination unit 12 in the image display apparatus according to the second embodiment includes a divided region feature determination unit 201, a multi-frame integration unit 102, and a determination speed adjustment unit 103.

  The divided region feature determination unit 201 determines the content feature determination value Si based on the combination content defined by the tint information value Ci (the tint feature value Ct) output from the tint information analysis unit 19, and integrates a plurality of frames. Output to the unit 102 and the determination speed adjustment unit 103, respectively.

  In the example shown in FIGS. 15 to 18, a predetermined number is obtained by the maximum (3 to the 18th power) combinations of 18 kinds of detailed divided regions Dd (1) to Dd (18) defined by the color feature value Ct. The divided region feature determination unit 201 determines the content feature determination value Si.

  Note that the combination details of content feature determination by the divided region feature determination unit 201 can be arbitrarily created based on the viewer's preference and the video database.

  In addition, the divided region feature determination unit 201 may perform determination based on a part of information of 18 kinds of combination contents defined by the tint information value Ci (the tint feature value Ct). For example, it is possible to select determination contents such as determining hue characteristics based only on the detailed division regions Dd (7) to Dd (12) (see FIG. 17). Thus, the hue feature determination speed is increased by reducing the amount of information used for the hue feature determination processing by the divided region feature determination unit 201 among the information defined by the hue information value Ci (color feature value Ct). , Can reduce the required memory capacity.

  The content feature determination unit 12 illustrated in FIG. 19 is an example of a configuration, and the content feature determination unit 12 may be configured to determine the content feature information value Si based on the tint information value Ci (the tint feature value Ct). Anything is acceptable.

  As described above, in the divided region feature determination unit 201 in the image display device according to the second embodiment, the color information value Ci (the color feature value Ct) obtained by the color information analysis unit 19 based on the image signal Db is defined. Based on information related to all of the hue information H, saturation information S, and lightness information V, the features are statistically classified and the content feature determination value Si is determined, thereby accurately determining the content feature and genre of the video. Can be determined.

  The processes of the multi-frame integration unit 102 and the determination speed adjustment unit 103 are the same as those of the multi-frame integration unit 102 and the determination speed adjustment unit 103 in the content feature determination unit 10 shown in FIG. The description will be omitted as appropriate.

  The image correction unit 6 (correction control unit 4, correction execution unit 5) according to the second embodiment is the same as that of the first embodiment and performs the same operation as that described in the first embodiment. The description will be omitted as appropriate.

  As described above, in the image display device according to the second embodiment, based on the content feature determination value Si for one frame determined from the color information value Ci, a plurality of content feature determination values based on a feature amount for a plurality of frames. Since Fi is determined, the characteristics of the content can be determined more accurately. Further, since the multi-frame integration unit 102 in the content feature determination unit 12 changes the content determination speed based on the determination speed information value Ti, the timing for determining the content is adjusted according to the feature of the content, and the image The timing for performing image correction by the correction unit 6 can be adjusted. As a result, it is possible to perform image correction in which a change in image quality is not noticeable.

  The image display apparatus according to the second embodiment is not limited to the video display apparatus and can be used in other fields related to video as a technique for more accurately determining the feature amount of the video content from the color information value Ci. For example, a video recording device such as a video recorder such as a hard disk or a DVD may be used as another field.

  On the other hand, in order to perform content feature determination using the technique described in Patent Document 2 described above, it is necessary to hold multimedia data whose content type is known in advance. . For this reason, classification cannot be performed for first-time programs.

  In Patent Document 3, since the video signals are classified based only on the luminance histogram based on the lightness signal V, they are classified using the hue information H, the saturation information S, and the lightness information V included in the image signal Db. Compared to the case, classification accuracy is inferior, and detailed and accurate classification is not possible.

  In the image display device according to the second embodiment, the color space is divided into a plurality of color spaces based on the HSV color space. For example, the RGB color space, the CMY color space, the HLS color space, the YCbCr color space, the Lab color system, The present invention can also be applied to all other color spaces and color systems such as the L * a * b * color system, the L * u * v * color system, and the like.

<Embodiment 3>
FIG. 20 is a block diagram showing a configuration of an image display apparatus according to Embodiment 3 of the present invention. The image display device according to the third embodiment includes the image processing device 27 in place of the image processing device 7 in the image processing device according to the first embodiment described above.

  The image processing apparatus 27 according to the third embodiment includes a color information analysis unit 29, a content feature determination unit 13, and an image correction unit 6. The image correction unit 6 includes a correction control unit 4 and a correction execution unit 5. is doing. The image signal Db obtained from the reception unit 2 is input to the color information analysis unit 29 and the correction execution unit 5 of the image processing device 27, respectively.

  The hue information analysis unit 29 detects the hue information (feature) value Ci in each frame from the hue information H, the saturation information S, and the lightness information V included in the image signal Db, and sends them to the content feature determination unit 13. Output.

  The content feature determination unit 13 determines the feature of the video content for one frame based on the color information value Ci, analyzes the video content for a plurality of frames, and then, as a result of the analysis, the multiple content feature determination value Fi is output to the correction control unit 4.

  The correction control unit 4 calculates a correction parameter Pi used when the correction execution unit 5 performs image correction on the image signal Db based on the multiple content feature information Fi, and outputs the correction parameter Pi to the correction execution unit 5.

  FIG. 21 is a block diagram illustrating details of an internal configuration example of the color information analysis unit 29 illustrated in FIG. 20. As shown in FIG. 21, the tint information analysis unit 29 in the image display apparatus according to the third embodiment includes a tint information detection unit 190c, a color space region division unit 291c, a feature information calculation unit 93c, and a feature information analysis unit. 292c.

  In the example illustrated in FIG. 21, the image signal Db obtained from the receiving unit 2 and including the hue information H, the saturation information S, and the lightness information V is input to the color information detection unit 190c.

  The hue information detection unit 190c extracts the hue information H, the saturation information S, and the lightness information V from the image signal Db for one frame, and outputs them to the color space region dividing unit 291c.

  The color space region dividing unit 291c, which is a division statistical processing unit, receives the hue information H, the saturation information S, and the lightness information V from the hue information detection unit 190c, and receives the division content instruction information Cb from the feature information analysis unit 292c. receive.

  Then, the color space region dividing unit 291c divides the color space into a plurality of color regions (first plurality of divided regions) based only on the brightness information V in the image signal Db for one frame. Then, the color space region dividing unit 291c performs the division statistical processing on the first plurality of divided regions from the color space to detect the region division information JDd which is the first statistical result, and the feature information calculating unit 93c. Output to. In the third embodiment, the HSV color space is used as an example of the color space.

  In the third embodiment, the color space region dividing unit 291c divides the color space into a plurality of color regions (first plurality of divided regions) based only on the brightness information V, but for one frame. Of course, the color space can be divided into a plurality of color regions based on any one of the hue information H, the saturation information S, and the lightness information V in the image signal Db.

  When the color space region dividing unit 291c receives the divided content instruction information Cb for instructing detailed division after outputting the region division information JDd as the first statistical result, the color space region dividing unit 291c further performs the following processing. That is, the color space region dividing unit 291c divides the color space into a plurality of color regions (second plurality of divided regions) based on at least one of the saturation information S and the hue information H in the image signal Db for one frame. To divide. Then, the color space region dividing unit 291c further performs the division statistical processing in the second plurality of divided regions from the color space to further detect the region division information JDd as the second statistical result, and the feature information calculating unit 93c Output to. The division content instruction information Cb includes the instruction content of specific instruction contents to be divided in detail (only the hue information H, a combination of the hue information H and the saturation information S, etc.).

  The feature information calculation unit 93c and the feature information analysis unit 292c are based on the region division information JDd (first or second statistical result) received from the color space region division unit 291c, and will be described later in detail. It functions as a divided region analysis processing unit that generates value accompanying information Ct2 and divided content instruction information Cb.

  The feature information calculation unit 93c extracts the region feature information Dds from the region division information JDd generated by the color space region division unit 291c, and outputs it to the feature information analysis unit 292c.

  The feature information analysis unit 292c calculates the color feature value Ct in the image signal Db for one frame based on the region feature information Dds generated by the feature information calculation unit 93c, and determines the content feature as the color information value Ci. To the unit 13. Therefore, when the region feature information Dds is based on the first statistical result based only on the lightness information V, a tint feature value Ct (first tint feature value) based on the lightness information V is obtained. On the other hand, when the region feature information Dds is based on the second statistical result including at least one of the hue information H and the saturation information S, the hue feature value Ct (the first feature value based on at least one of the hue information H and the saturation information S). 2 color characteristic values).

  Furthermore, the feature information analysis unit 292c outputs, to the content feature determination unit 13, the tint feature value accompanying information Ct2 that is supplementary information about the tint feature value Ct as necessary.

  When the color feature value Ct corresponds to the first color feature value, the feature information analysis unit 292c determines the value of the color feature value Ct, thereby determining the second plurality of divided regions. Whether detailed analysis is necessary or not is determined. If it is determined that detailed analysis is necessary for the second plurality of divided regions, the feature information analysis unit 292c outputs division content instruction information Cb for instructing detailed division to the color space region dividing unit 291c. . When the divided content instruction information Cb instructs detailed division, the divided content instruction information Cb includes the divided content.

  When the feature information analysis unit 292c determines that detailed analysis in the second plurality of divided regions is not necessary, the color feature value Ct (first color feature value) based on the lightness information V is directly used as the color. The taste information value Ci is output to the content feature determination unit 13.

  In the example shown in FIG. 21, when the input image signal Db is an interlace signal, the color information is converted by the color information detection unit 190 c into two frames as an image signal for one frame, and then the color space region The area division information JDd is obtained from the dividing unit 291c. As described above, the color information detection unit 190c may perform preprocessing for obtaining the region division information JDd.

  As another example of pre-processing, when the video signal Db is shown in a color system different from the hue information H, the saturation information S, and the lightness information V, such as an RGB signal, the hue information H is expressed by a known calculation formula. It is also conceivable to convert to saturation information S and lightness information V.

  FIG. 22 is a graph showing an example of the region division information JDd (first statistical result) in the first plurality of divided regions that is always generated by the color space region dividing unit 291c. In the example shown in FIG. 22, a brightness histogram related to the brightness information V is generated as the region division information JDd. Hereinafter, information defining a histogram of lightness information V in the image signal Db generated by the color space region dividing unit 291c is referred to as a lightness histogram Dbv.

  In the example shown in FIG. 22, the brightness information V is used to create the brightness histogram Dbv. However, even if the histogram is created using hue information H and saturation information S other than the brightness information V, the histogram is created. It doesn't matter. Further, the lightness histogram Dbv may be created by combining the hue information H, the saturation information S, and the lightness information V.

  In the third embodiment, the lightness information V is divided into a plurality of regions by creating a histogram relating to the lightness information V. However, the region is divided by a method other than creating a histogram. It is also possible.

  In the example shown in FIG. 22, the horizontal axis in the figure indicates the gradation value (class), and the vertical axis indicates the frequency, that is, the number of pixels with respect to the brightness of the image signal Db for one frame. In the following description, the lightness information V of the image signal Db is composed of, for example, 8-bit data, and the gradation value takes a value from “0” to “255”. 256 ". The image signal Db is represented by 8 bits in this example, but may be represented by other gradation numbers such as 10 bits and 6 bits.

  The color space area dividing unit 291c according to the third embodiment divides, for example, 256 gradations into 32 areas every 8 gradations, and the 32 areas are used as histogram classes. Then, a value near the center value in each class, in this example, an integer value closest to the center value and larger than that is used as the representative value of the class. For example, in a class composed of gradation values “0” to “7”, the central value is “3.5”, so the representative value of the class is “4”. The numbers on the horizontal axis in FIG. 22 indicate the representative values of each class. Hereinafter, the brightness of the brightness histogram whose representative value is “i” is referred to as Cbv (i), and the frequency thereof is referred to as Dbv (i).

  If the center value of the class is an integer, the center value may be used as the representative value of the class. Further, even when the center value of the class is not an integer but a decimal number as in this example, the center value of the class may be adopted as the representative value of the class. When the center value of a class is a decimal, the amount of calculation can be reduced by adopting an integer near the center value of the class as a representative value of the class as in this example.

  As described above, in the color space region dividing unit 291c according to the third embodiment, since the region including eight continuous gradation values is set as one class, each frequency of the histogram shown in FIG. Is the sum of the signals. For example, the frequency indicated by the numerical value 4 on the horizontal axis corresponds to the sum of the signals from the gradation value 0 to the gradation value 7 included in the luminance signal Db for one frame.

  Note that, unlike the histogram of FIG. 22, the histogram may be generated by counting the frequency for each gradation value. That is, each class may be configured with one gradation value. In this case, the representative value of each class is the gradation value itself constituting the class.

  Further, when the number of gradations is divided, the number of divisions may be set to other than 32, and the division number and range of the histogram may be set freely. For example, when the number of gradations is “256”, the gradation values “0” to “32” and “200” to “255” can be in four gradations, and the others can be in 24. In this way, by changing the number of divisions, it is possible to adjust the amount of calculation in the color space region dividing unit 291c and the accuracy of the histogram.

  In the histogram generated by the color space region dividing unit 291c, the feature information calculation unit 93c accumulates the frequencies from the maximum to the minimum of the class, and the accumulated frequency HYW obtained thereby becomes larger than the predetermined threshold YA for the first time. Extract the representative value of the class. Then, the extracted value is calculated as the maximum gradation information value Vimax.

  On the other hand, in the histogram generated by the color space region dividing unit 291c, the frequency is accumulated from the minimum to the maximum of the class, and the representative value of the class in which the cumulative frequency HYB obtained thereby becomes larger than the predetermined threshold YB for the first time is obtained. Extract. Then, the extracted value is calculated as the minimum gradation information value Vimin.

  Further, in the histogram generated by the color space region dividing unit 291c, the frequencies are accumulated from the minimum to the maximum of the class, and the obtained cumulative frequency HYB is more than a predetermined threshold YC (for example, half of the total number of pixels). Extracts the representative value of the class that grows for the first time Then, the extracted value is calculated as the intermediate luminance gradation Vimid. Note that the cumulative frequency HYW may be used to calculate the intermediate luminance gradation Vimid.

  In the histogram shown in FIG. 22, the representative value of the class in which the cumulative frequency HYW becomes larger than the threshold value YA for the first time is “188”, so “188” is the maximum gradation information value Vimax. Thus, unlike the maximum gradation value “204” in the video signal Db for one frame, the maximum gradation information value Vimax is a value according to the maximum gradation value detected using the cumulative frequency HYW and the threshold value YA. is there. Note that the maximum gradation value itself may be used as the maximum gradation information value in order to reduce processing. The same applies to the following minimum gradation information values.

  In the example of FIG. 22, the representative value of the class in which the cumulative frequency HYB becomes larger than the threshold value YB for the first time is “20”, so “20” is the minimum gradation information value Vimin. This minimum gradation information value Vimin is not the minimum gradation value in the image signal Db for one frame, but a value according to the minimum gradation value detected using the cumulative frequency HYB and the threshold value YB.

  Since the representative value of the class in which the cumulative frequency HYB becomes larger than the threshold value YC for the first time is “76”, this “76” becomes the intermediate gradation information value Vimid. Normally, this intermediate gradation information value Vimid is a gradation value when it reaches half (50%) of the total number of pixels in the image signal Db for one frame.

  Further, the feature information calculation unit 93c calculates an average lightness information value in the lightness information V for one frame from the lightness information V of the image signal Db for one frame obtained through the color space region dividing unit 291c. This is calculated as an average gradation information value Viave. More specifically, the lightness signal gradation number is Vi, and the pixel number of the lightness signal gradation is Vdi.

(Average brightness information value) = Σ (Vi × Vdi) / ΣVdi (1)
This average brightness information value is output as an average gradation information value Viave.

  In this way, the feature information calculation unit 93c is a first statistical result area that defines the minimum gradation information value Vimin, the maximum gradation information value Vimax, the intermediate gradation information value Vimid, and the average gradation information value Viave. The feature information Dds is output. As described above, since information values such as the lightness information value V and the maximum gradation information value Vimax statistically classified from the lightness histogram Dbv are included, the lightness information value with high accuracy classified statistically for each frame. Can be obtained.

  In the above example, the color space region dividing unit 291c divides the histogram evenly. However, the histogram can be divided unevenly so that the range of gradation values to be counted can be freely set. You may comprise. Thereby, the amount of calculation can be reduced.

  For example, it is assumed that the minimum brightness information value condition (threshold YB) is set to 16 gradations and the maximum brightness information value condition (threshold YW) is set to 160 gradations. In the case of equal division, it is possible to divide into 16 areas by dividing every 16 gradations so as to match the minimum brightness information value condition. On the other hand, if it can be divided non-uniformly, the minimum brightness information value condition and the maximum brightness information condition value described above are satisfied with at least three area divisions of 0 to 16 gradations, 16 to 160 gradations, and 160 to 255 gradations. In this way, it is possible to significantly reduce the amount of calculation associated with the reduction in the number of divisions.

  Note that, even if the minimum brightness information value condition and the maximum brightness information value condition are not met, the minimum brightness information value condition and the maximum brightness information value condition may be estimated by arithmetic operation from values of other gradations. However, in that case, accuracy decreases. For example, when obtaining gradation values of 16 gradations, a method of estimating gradation values of 16 gradations by obtaining gradation values of 12 gradations and 20 gradations and taking an average value may be considered. This example may be used when a matching gradation value cannot be calculated due to restrictions such as hardware and software.

  The feature information analysis unit 292c has each frequency of the area feature information Dds (maximum gradation information value Vimax, intermediate gradation information value Vimid, minimum gradation information value Vimin, average gradation information value Viave) extracted from the brightness histogram Dbv. Then, it is determined whether the value is smaller than the predetermined brightness determination threshold value TDbv1, is greater than the threshold value TDbv1, and is equal to or smaller than the threshold value TDbv2, which is a predetermined threshold value, or greater than the brightness determination threshold value TDbv2.

  The feature information analysis unit 292c then includes three classification information values “small (low)” for each of the maximum gradation information value Vimax, the intermediate gradation information value Vimid, the minimum gradation information value Vimin, and the average gradation information value Viave. The tint feature value Ct that defines the lightness determination information value Dbvc indicating any one of “”, “medium”, and “large (high)” is determined.

  FIG. 23 is an explanatory diagram illustrating an example of generation of the brightness determination information value Dbvc by the feature information analysis unit 292c. As shown in the figure, the maximum gradation information value Vimax is a value between a predetermined threshold value TDbv1 and a TDbv2 value that is a predetermined threshold value larger than the threshold value TDbv1, and thus the lightness determination information value Dbvc is the maximum gradation value. It is defined that the information value Vimax is “medium”.

  In this example, the predetermined threshold value is classified into two types and the classification information value is classified into three types of large, medium and small, but may be classified into other than two types and three types. That is, the threshold value may be more or less than two types. More detailed classification is possible by adjusting a large number of threshold values. Further, the processing amount can be reduced by reducing the number of thresholds. Similarly, there may be more or less classification information values.

  Further, the threshold value may be changed for each parameter in the region feature information Dds. For example, the maximum gradation information value Vimax that is one parameter is determined only by a predetermined threshold TDbv1, and the minimum gradation information value Vimin that is another parameter is determined by a predetermined threshold TDbv1 and another predetermined threshold TDbv2. , TDbv3 can also be used to obtain the lightness determination information value Dbvc. In this way, by changing the threshold value for each parameter of the region feature information Dds, the lightness region desired to be analyzed in detail is focused on, and the analysis accuracy of the lightness region that does not require detailed analysis is reduced. The effect of can be obtained.

  Prior to outputting the color feature value Ct to the content feature determination unit 13 as the color information value Ci, the feature information analysis unit 292c further determines the second plurality of color feature values Ct based on the obtained color feature value Ct. Divided content instruction information Cb that indicates whether detailed analysis is required or not necessary based on the area division information Dd in the divided area is calculated.

  FIG. 24 is an explanatory diagram showing the relationship between the color feature value Ct (first color feature value) and the divided content instruction information Cb by the feature information analysis unit 292c in a table format. As shown in the figure, the feature information analysis unit 292c, in principle, uses the color feature value Ct when the intermediate gradation information value Vimid indicated by the color feature value Ct is “large” (first case). Is not output to the content feature determination unit 13 as the color information value Ci.

  In the first case, the feature information analysis unit 292c outputs division content instruction information Cb instructing detailed division related to the hue information H and the saturation information S to the color space region division unit 291c, and the hue information H and the saturation are output. The color space region dividing unit 291c is caused to execute the division statistical processing for the information S in the second plurality of divided regions.

  On the other hand, when the intermediate tone information value Vimid indicated by the color feature value Ct is “medium” (second case), the feature information analysis unit 292c performs the color feature value Ct (first color feature). Value) is not output to the content feature determination unit 13 as a color information value Ci in principle.

  In the second case, the feature information analysis unit 292c outputs division content instruction information Cb instructing detailed division related to the hue information H to the color space region division unit 291c, and the second plurality of divisions related to the hue information H. The color space region dividing unit 291c is caused to execute division statistical processing in the region.

  In the first and second cases described above, the feature information analysis unit 292c eventually uses the region feature information Dds obtained again from the feature information calculation unit 93c based on the region division information JDd that is the second statistical result. Based on this, the color feature value Ct (second color feature value) is obtained again, and information including at least the color feature value Ct is output to the content feature determination unit 13 as the color information value Ci.

  Also, the feature information analysis unit 292c, when the intermediate gradation information value Vimid indicated by the color feature value Ct is “small” (third case), the color feature value Ct (first color tone). (Feature value) is directly output to the content feature determination unit 13 as the color information value Ci. In the third case, the feature information analysis unit 292c does not output the division content instruction information Cb that instructs detailed division.

  The determination of the instruction content of the divided content instruction information Cb by the feature information analysis unit 292c in the first to third cases is uniquely given for each lightness determination information value Dbvc in the area feature information Dds. Note that the instruction content of the divided content instruction information Cb may be changed based on the combination of the lightness determination information value Dbvc in the color feature value Ct.

  In the example shown in FIG. 24, in the first case where the intermediate gradation information value Vimid is “large”, the color space region dividing unit 291c performs the second processing related to “hue information H / saturation information S”. Division statistical processing in a plurality of divided regions is performed again. As a result, the color space region dividing unit 291c performs the saturation information S and the frequencies of the second plurality of divided regions based on the saturation information S, that is, the second plurality of divided regions of the image signal Db for one frame. The number of pixels is calculated, and the area division information JDd (H, S, V, D), which is information including the frequency, is output to the feature information calculation section 93c together with the hue information H and the saturation information S of the corresponding divided area. This area division information JDd becomes the second statistical result.

  This is because the hue information H and the saturation information S obtained from the second statistical result are judged by only the lightness information V when the intermediate gradation information value Vimid obtained from the first statistical result is “large”. This is because the accuracy is improved by making a judgment including the above. For example, when the intermediate gradation information value Vimid is “large”, shades of green, red, and the like are included as colors that are often included, but it is difficult to distinguish these shades from the brightness histogram Dbv. For this reason, it is necessary to output information including the hue information H and the saturation information S from the color space region dividing unit 291c in order to perform detailed analysis on the corresponding region or the entire region.

  As a result, the feature information calculation unit 93c and the feature information analysis unit 292c cause the color information value Ci to include at least the color feature value Ct (second color feature value) based on the hue information H and the saturation information S. Is output to the content feature determination unit 13.

  In the detailed analysis by the content feature determination unit 13, the content that can be determined from the value of the color information value Ci corresponding to the first color feature value may be reflected in the color feature value associated information Ct2. Absent. For example, in the case of the first case where the intermediate gradation information value Vimid is “large”, the content feature information value Si obtained in the content feature determination unit 13 is sport (S1) or drama (S5). Probability is high. Therefore, the feature information analysis unit 292c outputs to the content feature determination unit 13 the color feature value accompanying information Ct2 that indicates that there is a high possibility of being a sport (S1) or a drama (S5). Therefore, when the content feature determination unit 13 performs a detailed analysis on the hue information value Ci corresponding to the second hue feature value based on the hue information H and the saturation information S, the content feature value accompanying information Considering the information of Ct2, the choice of music (S2), studio (S3), movie (S4) is removed, and the number of processing is reduced by determining whether it is either sports (S1) or drama (S5) Is possible.

  When the feature information analysis unit 292c determines that detailed analysis is necessary as described above, the feature information analysis unit 292c outputs division content instruction information Cb instructing detailed division. In this case, the color space region dividing unit 291c performs detailed region dividing processing in the second plurality of divided regions based on the instruction content of the divided content instruction information Cb from the feature information analyzing unit 292c.

  On the other hand, if the feature information analysis unit 292c determines that it is not necessary to perform detailed analysis as described above, the color feature value Ct that is the first color feature value is used as the color information value Ci as it is as the content feature. Output to the determination unit 13.

  In the third embodiment, the detailed analysis based on the detailed division instruction of the divided content instruction information Cb from the feature information analysis unit 292c is performed only once. However, the analysis area is changed as necessary to change the details. The analysis may be performed a plurality of times. Further, when performing a detailed analysis, the number of divisions of the region and the setting of the threshold value may be changed as necessary. As a result, detailed analysis that cannot be obtained in the previous analysis becomes possible.

  An example of detailed analysis using the second plurality of divided regions is shown in FIG. In the example shown in FIG. 25, detailed analysis regarding the hue information H and the saturation information S is performed on the image signal Db, and the color space region dividing unit 291c performs 3 analysis based on the hue information H and the saturation information S. The division is performed by a plurality of divided areas (second plurality of divided areas) composed of the types of divided areas Dd (19), Dd (20), and Dd (21). In this example, Dd (19) is a color that is not vivid and is close to black or white in the HSV color space, Dd (20) is a color close to green, and Dd (21) is a color close to red. ing.

  In this way, the color space region dividing unit 291c, when the division content instruction information Cb instructs detailed division, the second color space divided based on at least one of the hue information H and the saturation information S. Division statistical processing in a plurality of divided regions is performed, and region division information JDd as a second statistical result is output.

  These second plurality of divided regions are also classified into one of “large (high)”, “medium”, and “small (low)” according to the threshold, as shown in the example in the feature information analysis unit 292c, and detailed division is performed. A combination of the hue / saturation determination information value Ddshc (i) in the regions Dd (19), Dd (20), and Dd (21) is set as a color feature value Ct (second color feature value), and if necessary The color information value Ci including the content of the first color specific value is output. In this way, by performing detailed analysis, it is possible to accurately determine the distinction between genres that have values close to those represented by the brightness information V.

  Another example of detailed analysis is to perform detailed analysis on only a part of the region. For example, as shown in FIG. 26, a method of performing a detailed analysis on the hue information H and the saturation information S with respect to the region Dbv (100) where the frequency is maximum in the brightness histogram Dbv can be considered. In this case, the division content instruction information Cb includes lightness information V having the highest frequency in the lightness histogram Dbv, and information instructing detailed division related to the hue information H and the saturation information S.

  As another example, a method of performing a detailed analysis on the hue information H and the saturation information S for a region between the minimum gradation information value Vimin and the maximum gradation information value Vimax is also conceivable. Areas with high and low brightness information V have a color close to black and white. For this reason, for the regions with high and low brightness information V, detailed analysis is not performed using only the first plurality of divided regions, and the second plurality of regions are further used for the other regions. By performing detailed analysis, it is possible to reduce the amount of calculation required. Thus, by performing detailed analysis only in a part of the region, it is possible to omit extra analysis and reduce the amount of calculation.

  In the third embodiment, the detailed analysis on the hue information H or the saturation information S is performed based on the analysis result (first color characteristic value) on the lightness information V. The detailed analysis is performed on the hue information H, Any information may be used as long as at least one of the saturation information S and the lightness information V is used. In addition, as an area for performing detailed analysis, any of existing divided areas, newly created divided areas, and the entire HSV hue space may be used.

  FIG. 27 is an example of a block diagram illustrating a detailed configuration of the content feature determination unit 13. As shown in the figure, the content feature determination unit 13 in the image display apparatus according to the third embodiment includes a feature determination unit 301 such as brightness, a multi-frame integration unit 102, and a determination speed adjustment unit 103.

  The feature determination unit 301 such as brightness determines the content feature determination value Si based on the combination content defined by the tint information value Ci (the tint feature value Ct) output from the tint information analysis unit 29, and integrates a plurality of frames. Output to the unit 102 and the determination speed adjustment unit 103. At this time, when useful instruction content exists in the color feature value accompanying information Ct2, the content feature determination value Si is determined in consideration of the content.

  For example, in the example of FIG. 24, when it corresponds to the first case where the intermediate gradation information value Vimid is “large”, as described above, there is a high possibility of being a sport (S1) or a drama (S5). Since the color feature value accompanying information Ct2 is received from the color information analysis unit 29, the lightness feature determination unit 301 can make a determination narrowed down to sports (S1) or drama (S5).

  When the color information value Ci (color characteristic value Ct) corresponds to the first color characteristic value based on the first statistical result, the feature determination unit 301 such as brightness adjusts the content characteristic based on the color information value Ci. A determination value Si is determined.

  In the example shown in FIG. 22 to FIG. 24, there are a maximum of 81 (3 to the 4th power) combinations based on four types of maximum gradation information values Vimax defined by the color information value Ci (color characteristic value Ct). A predetermined number of content feature determination values Si are determined by the feature determination unit 301 such as brightness.

  On the other hand, if the color information value Ci (color characteristic value Ct) corresponds to the second color characteristic value based on the second statistical result, the lightness characteristic determination unit 301 sets the color information value Ci to the color information value Ci. In addition, the content feature determination value Si is determined in consideration of the color feature value accompanying information Ct2 as necessary.

  Note that the content feature determination combination content based on the color information value Ci corresponding to the first or second color feature value by the feature determination unit 301 such as brightness can be arbitrarily set based on the viewer's preference and the video database. Can be created.

  The feature determination unit 301 such as lightness may be determined based on some information of the combination content defined by the color information value Ci.

  The content feature determination unit 13 illustrated in FIG. 27 is a configuration example, and the content feature determination unit 13 may be configured to determine the content feature information value Si based on the tint information value Ci (the tint feature value Ct). Anything is acceptable.

  Note that the processing of the multi-frame integration unit 102 and the determination speed adjustment unit 103 is the same as that of the multi-frame integration unit 102 and the determination speed adjustment unit 103 in the content feature determination unit 10 in Embodiment 1 shown in FIG. The same reference numerals are given and description thereof is omitted as appropriate.

  The image correction unit 6 according to the third embodiment is the same as that of the first embodiment shown in FIG. 1 and performs the same operation as that described in the first embodiment. Omitted as appropriate.

  In the image display device according to the third embodiment, the color space region dividing unit 291c, which is the division statistical processing unit of the color information analyzing unit 29, performs division related to color information (lightness information V) in the first plurality of divided regions. Statistical processing is performed to obtain a first statistical result.

  On the other hand, in the feature information calculation unit 93c and the feature information analysis unit 292c in the feature information analysis unit 292c constituting the divided region analysis processing unit, the color feature value Ct obtained based on the first statistical result satisfies a predetermined requirement. In this case, the division content instruction information Cb for instructing the detailed division is output to the color space region dividing unit 291c.

  The color space area dividing unit 291c performs division statistical processing on the color information (hue information H, saturation information S) in the second plurality of divided areas in response to the detailed division instruction of the divided content instruction information Cb. To obtain a second statistical result.

  Therefore, the image display apparatus according to the third embodiment has an image correction process that can be performed in a relatively short time using only the color feature value Ct based on the first statistical result, and a color based on at least the second statistical result. Since image correction processing with relatively high accuracy by the feature value Ct can be selectively performed, more appropriate image correction can be performed according to the state of the image signal.

  In addition, in the image display device according to the third embodiment, based on the content feature determination value Si for one frame determined from the color information value Ci, the multiple content feature determination value Fi is obtained from the feature amount for a plurality of frames. Since the determination is made, the characteristics of the content can be determined more accurately. In addition, since the multi-frame integration unit 102 in the content feature determination unit 13 changes the content determination speed based on the determination speed information value Ti, the timing for determining the content is adjusted according to the feature of the content, The timing for performing image correction by the image correction unit 6 can be adjusted. As a result, it is possible to perform image correction in which a change in image quality is not noticeable.

  The image display device according to the third embodiment is not limited to the video display device and can be used in other fields related to video as a method for more accurately determining the feature amount of the video content from the color information value. For example, a video recording device such as a video recorder such as a hard disk or a DVD may be used as another field.

  On the other hand, in order to perform content feature determination using the technique described in Patent Document 2 described above, it is necessary to hold multimedia data whose content type is known in advance. . For this reason, classification cannot be performed for first-time programs.

  In Patent Document 3, since the video signals are classified based only on the luminance histogram based on the lightness signal V, they are classified using the hue information H, the saturation information S, and the lightness information V included in the image signal Db. Compared to the case, classification accuracy is inferior, and detailed and accurate classification is not possible.

  In the image display device according to the third embodiment, the color space is divided into a plurality of color spaces based on the HSV color space. For example, the RGB color space, the CMY color space, the HLS color space, the YCbCr color space, the Lab color system, The present invention can also be applied to all other color spaces and color systems such as the L * a * b * color system, the L * u * v * color system, and the like.

<Application to program>
Corresponding to Embodiment 1 capable of computer-executing at least a part of the constituent parts (processing by the tint information analysis unit 9, the content feature determination unit 10 and the correction control unit 4 in FIG. 1) in the image display apparatus of Embodiment 1 Can be described as a program.

  Similarly, an embodiment in which at least a part of the components in the image display apparatus according to the second embodiment (processing by the color information analysis unit 19, the content feature determination unit 12, and the correction control unit 4 in FIG. 13) can be executed by a computer. It can be described as a program corresponding to form 2.

  Similarly, at least a part of the constituent parts (processing by the color information analysis unit 29, the content feature determination unit 13 and the correction control unit 4 in FIG. 20) in the image display apparatus according to the third embodiment can be executed by a computer. It can be described as a program corresponding to form 3.

  Therefore, by causing the computer to execute the above-described programs corresponding to the first to third embodiments, in each embodiment, there is an effect that the characteristics of the content obtained from the image signal can be accurately determined.

1 is a block diagram showing a configuration of an image display apparatus according to Embodiment 1 of the present invention. It is a block diagram which shows the structure of the color information analysis part shown in FIG. It is explanatory drawing which shows the hue histogram produced | generated by the color space area | region division part shown in FIG. It is explanatory drawing which shows the threshold value condition with respect to the hue histogram shown in FIG. It is a block diagram which shows the structure of the content characteristic determination part shown in FIG. It is explanatory drawing which shows the content feature determination content by the hue feature determination part shown in FIG. 5 in a table format. It is explanatory drawing which shows the content feature determination content by the hue feature determination part shown in FIG. 5 in a table format. FIG. 6 is an explanatory diagram (part 1) illustrating an example of a plurality of content feature information values determined by a plurality of frame integration units illustrated in FIG. 5. FIG. 6 is an explanatory diagram (part 2) illustrating an example of a plurality of content feature information values determined by a plurality of frame integration units illustrated in FIG. 5. FIG. 6 is an explanatory diagram (part 3) illustrating an example of a plurality of content feature information values determined by a plurality of frame integration units illustrated in FIG. 5. It is explanatory drawing showing the example of adjustment content of the judgment speed information value produced | generated by the judgment speed adjustment part shown in FIG. It is explanatory drawing for demonstrating operation | movement of the correction control part shown in FIG. It is a block diagram which shows the structure of the image display apparatus which is Embodiment 2 of this invention. It is a block diagram which shows the structure of the color information analysis part shown in FIG. It is a figure which shows an example of the area | region division in the HSV color space produced | generated by the color space area | region division part shown in FIG. It is explanatory drawing which shows an example of the area | region division in the HSV color space produced | generated by the color space area | region division part shown in FIG. It is explanatory drawing which shows an example of the area | region division in the HSV color space produced | generated by the color space area | region division part shown in FIG. It is explanatory drawing which shows an example of the area | region division in the HSV color space produced | generated by the color space area | region division part shown in FIG. It is a block diagram which shows the structure of the content characteristic determination part shown in FIG. It is a block diagram which shows the structure of the image display apparatus which is Embodiment 3 of this invention. It is a block diagram which shows the structure of the color information analysis part shown in FIG. It is explanatory drawing which shows the brightness histogram obtained by equally dividing | segmenting produced | generated by the color space area | region division part shown in FIG. It is explanatory drawing which shows the lightness histogram and threshold value condition which were equally divided | segmented produced | generated by the color space area | region division part shown in FIG. It is explanatory drawing which shows the output content of the division content instruction | indication information produced | generated by the feature information analysis part shown in FIG. 21 in a table format. It is explanatory drawing which shows an example of the detailed division area performed in the color space area division part shown in FIG. It is explanatory drawing which shows the example of the detailed analysis performed in the color information analysis part shown in FIG. It is a block diagram which shows the structure of the content characteristic determination part shown in FIG.

Explanation of symbols

  2 receiving unit, 4 correction control unit, 5 correction executing unit, 6 image correcting unit, 7, 17, 27 image processing device, 8 display unit, 9, 19, 29 color information analyzing unit, 10, 12, 13 content feature Determination unit, 90c Hue detection unit, 91c, 191c, 291c Color space region division unit, 92c Hue information analysis unit, 292c Feature information analysis unit, 101 Hue feature determination unit, 102 Multiple frame integration unit, 103 Judgment speed adjustment unit, 190c Color information detection unit, 192c Color information analysis unit, 201 Divided region feature determination unit, 93c Feature information calculation unit, 301 Brightness etc. feature determination unit

Claims (6)

A color information analysis unit that receives an image signal and performs analysis processing on a plurality of divided areas divided in the color space for the color information obtained from the image signal for one frame to obtain a color information value The color information includes at least one of saturation information and hue information;
A content feature determination unit that determines content features for the image signal based on the color information value and obtains a determined content feature determination value;
An image correction unit that performs image correction on an image signal for one frame based on the determined content feature determination value;
Image display device. The image display device according to claim 1,
The content feature determination unit
A color feature determination unit that performs content feature determination on the image signal for one frame based on the color information value and obtains a temporary content feature determination value;
A plurality of frame analysis units for analyzing the provisional content feature determination value by an amount corresponding to a predetermined frame to obtain the determined content feature determination value;
A determination speed adjustment unit capable of variably setting a determination speed information value, and a determination speed for obtaining the determined content feature determination value by the plurality of frame analysis units is determined based on the determination speed information value.
Image display device. The image display device according to claim 2,
The determination speed adjustment unit variably sets the determination speed information value based on at least one of the temporary content feature determination value and the determined content feature determination value.
Image display device. The image display device according to any one of claims 1 to 3,
The color information analysis unit
A color information detecting unit for detecting the color information obtained from the image signal for one frame;
A divided statistical processing unit for performing statistical processing on the plurality of divided areas and obtaining a statistical result for the color information;
A divided region analysis processing unit that obtains the color information value based on the statistical result by the divided statistical processing unit,
Image display device. The image display device according to any one of claims 1 to 4, wherein:
The hue information includes a combination of at least two of lightness information, saturation information, and hue information.
Image display device. A color information analysis unit that receives an image signal and performs analysis processing on a plurality of divided areas divided in the color space for the color information obtained from the image signal for one frame to obtain a color information value The color information includes at least one of saturation information, hue information, and brightness information,
A content feature determination unit that determines content features for the image signal based on the color information value and obtains a determined content feature determination value;
An image correction unit that performs image correction on an image signal for one frame based on the determined content feature determination value;
The plurality of divided regions include at least a first and a second plurality of divided regions having different division contents.
The color information analysis unit
A color information detecting unit for detecting the color information obtained from the image signal for one frame;
A division statistical processing unit that performs division statistical processing on the color information in the first plurality of divided regions and outputs a first statistical result, wherein the division statistical processing unit receives division content instruction information, When the division content instruction information indicates detailed division, the division statistical processing is performed on the color information in the second plurality of divided areas, and a second statistical result is further output,
The color information value is calculated based on at least one of the first and second statistical results obtained by the divided statistical processing unit, and the divided content instruction information is output based on the first statistical result. The divided area analysis processing unit further includes: a divided content instructing the detailed division when the color information value obtained based on the first statistical result satisfies a predetermined requirement Output instruction information,
Image display device.

Images