JP5627507B2 - Moving picture encoding apparatus, moving picture decoding apparatus, moving picture encoding method, moving picture decoding method, and program - Google Patents

Description

  The present invention relates to a video encoding device, a video decoding device, a video encoding method, a video decoding method, and a program.

  Conventionally, in video compression encoding processing, a technique for improving image quality by performing filter processing on an encoded image has been proposed (for example, see Non-Patent Documents 1, 2, and 3).

  Non-Patent Document 1 discloses a standard for video compression coding. In this standard, it is possible to apply a filter process that reduces deterioration of block boundaries caused by compression encoding.

  Non-Patent Document 2 discloses a method of adaptively updating a filter for restoring coding degradation for each frame. In this method, filter coefficients are calculated for the entire screen.

  Non-Patent Document 3 discloses a technique of dividing a screen into two predetermined fixed rectangular areas and adaptively calculating filter coefficients for each area for a filter that restores encoding degradation.

Joint Video Team (JVT) of ISO / IEC MPEG and ITU-T VCEG, "Text of ISO / IEC 14496-10 Advanced Video Coding," July 2004. D.-H. Kim, H.-Y. Oh, O. Urhan, S. Erturk and T.-G. Chang, "Optimal Post-Process / In-Loop Filtering for Improved Video Compression Performance," IEEE Trans. Consumer Electronics, Vol.53, No.4, 2007. I. Lim, D. Park and C. Kim, "A Loop Filter with Segmentation Mode in Video Coding," ICIP, pp. 2073-2076, 2010.

  However, in the standard shown in Non-Patent Document 1, since the above-described filter processing is applied only to the block boundary, it is difficult to obtain a sufficient image quality improvement effect by the filter processing.

  In the method disclosed in Non-Patent Document 2, adaptive filter processing is performed on the entire screen. For this reason, the filter coefficients used in the filter process are only suitable for the entire screen on average, and the local optimum is poor, so the effect of improving the image quality by the filter process was limited. .

  In the method disclosed in Non-Patent Document 3, adaptive filter processing is performed for each divided region. However, since the shape of the region is fixed by two predetermined rectangles, the adaptability of the filter may not sufficiently follow the pattern. Therefore, it is conceivable to increase the number of areas. In this case, however, there is a problem that the processing amount is remarkably increased.

  Accordingly, the present invention has been made in view of the above-described problems, and provides a moving image encoding device, a moving image decoding device, a moving image encoding method, a moving image decoding method, and a program that can improve compression encoding performance. The purpose is to provide.

  The present invention proposes the following matters in order to solve the above problems.

  (1) The present invention allows a moving image coding apparatus (for example, a moving image coding apparatus AA in FIG. 1 or a later-described moving picture coding apparatus that permits adaptive filter processing and controls filter processing for each filter processing unit block in the coding process. A video encoding device CC), and based on the feature quantity of the image to be encoded (for example, corresponding to the original image a in FIG. 1), the image is classified into a plurality of types (for example, N types described later). ) For the feature quantity analyzing means (e.g., equivalent to the feature quantity analysis unit 5 in FIG. 2 or the feature quantity analysis unit 5A in FIG. 5) and adaptive filter coefficients for each of the plurality of types of areas. Using the filter coefficient calculation means to be obtained (for example, equivalent to the filter coefficient calculation unit 6 in FIG. 1) and the adaptive filter coefficients obtained by the filter coefficient calculation means for each of the plurality of types of regions, Adaptive filter processing means (for example, equivalent to the adaptive filter processing section 7 in FIG. 1) for performing filter processing on (for example, equivalent to the local decoded image i in FIG. 1), and all of the filter coefficient calculation means obtained by the filter coefficient calculation means Filter application information (e.g., corresponding to N types of adaptive filter coefficients j in FIG. 1) and filter application information (for each region) that is used by the adaptive filter processing means among all the adaptive filter coefficients. For example, the encoded data generating means (for example, the entropy encoding unit 4 in FIG. 1) generates encoded data (for example, corresponding to the encoded data c in FIG. 1). A video encoding apparatus characterized by comprising:

  According to the present invention, in a moving image encoding apparatus that allows adaptive filter processing and controls filter processing for each filter processing unit block in encoding processing, a feature amount analyzing unit, a filter coefficient calculating unit, an adaptive filter processing unit And an encoded data generating means. Then, the feature amount analysis means divides the image into a plurality of types of regions based on the feature amount of the image to be encoded, and the filter coefficient calculation means obtains adaptive filter coefficients for each of the plurality of types of regions. did. Further, the adaptive filter processing means performs the filtering process on the locally decoded image using the adaptive filter coefficient obtained by the filter coefficient calculation means for each of the plurality of types of regions. Further, all the adaptive filter coefficients obtained by the filter coefficient calculating means by the encoded data generating means, and filter application information indicating, for each region, those used by the adaptive filter processing means among all the adaptive filter coefficients; The encoded data including is generated.

  For this reason, the moving image encoding apparatus of (1) can divide an image into a plurality of types of regions based on the feature amount of the image to be encoded, and obtain an adaptive filter coefficient for each region. Therefore, it is possible to divide the image according to the design and obtain the adaptive filter coefficient suitable for the design for each divided region. Therefore, an adaptive filter coefficient that can follow the pattern while suppressing an increase in the processing amount can be obtained, and the compression encoding performance can be improved.

  (2) In the moving image encoding apparatus according to (1), the feature amount analyzing unit generates a plurality of predetermined images (for example, corresponding to the original image a in FIG. 1) to be encoded (for example, , Each of the plurality of regions is obtained by feature analysis, and the feature amount is classified into a plurality of types (for example, equivalent to N types described later), Proposing a moving picture coding apparatus characterized by classifying areas corresponding to feature quantities classified into the same type as areas of the same type and dividing the image into a plurality of types of classified areas. .

  According to the present invention, the feature amount analyzing means divides the image to be encoded into a plurality of predetermined regions, the feature amount is obtained by feature analysis for each of the plurality of regions, and the feature amounts are classified into a plurality of types. It was decided to classify. Then, the regions corresponding to the feature quantities classified into the same type are classified as the same type of region, and the image is divided into a plurality of classified regions.

  For this reason, the moving image encoding apparatus of (2) can divide the image to be encoded into a plurality of types of regions based on the feature amount.

  (3) The present invention relates to the moving picture encoding apparatus according to (2), wherein the feature amount analyzing means performs the feature for each of the plurality of predetermined regions (for example, corresponding to M described later). Edge strength calculating means for obtaining edge strength as a quantity (for example, equivalent to the edge strength calculating section 51 in FIG. 2), and edge strength obtained by the edge strength calculating means are represented by a plurality of threshold values (for example, (N -1) edge strength threshold processing means (e.g., equivalent to the edge strength threshold processing unit 52 in FIG. 2) that classifies into a plurality of types (e.g., equivalent to N types described later) using Edge strength base divided region generating means (for example, edge strength base divided region in FIG. 2) for classifying the image into a plurality of types (for example, equivalent to N types described later) based on the classification result by the edge strength classifying unit. And equivalent) to the generating section 53, have proposed a moving picture coding apparatus comprising: a.

  According to the present invention, the feature amount analyzing means is provided with edge strength calculating means, edge strength threshold processing means, and edge strength base divided region generating means. Then, the edge strength is calculated as a feature amount for each of a plurality of predetermined regions by the edge strength calculation means. Also, the edge strength threshold processing means classifies the edge strength obtained by the edge strength calculation means into a plurality of types using a plurality of types of threshold values. In addition, the image is classified into a plurality of types based on the classification result by the edge strength classification unit by the edge strength base divided region generation unit.

  For this reason, the moving image encoding apparatus of (3) can divide the image to be encoded into a plurality of types of regions using the edge intensity as the feature amount.

  (4) In the moving image encoding apparatus according to (2), the feature amount analyzing unit may include the feature for each of the plurality of predetermined regions (for example, corresponding to M described later). Complexity calculation means for obtaining the complexity of pixel values as quantities (for example, equivalent to the complexity degree calculation unit 54 in FIG. 5), and the complexity of pixel values obtained by the complexity degree calculation means are represented by a plurality of threshold values ( For example, complexity level threshold processing means (for example, complexity level threshold processing unit 55 in FIG. 5) classifies into a plurality of types (for example, corresponding to N types described later) using (N-1) threshold levels described later). And a degree-of-complexity-based divided region generation means (for example, the complexity shown in FIG. 5) that classifies the image into a plurality of types (for example, equivalent to N types described later) based on the classification result by the complexity level threshold processing means. Degree Proposes a video encoding device, characterized in that it comprises a corresponding) to scan the divided area generator 56.

  According to this invention, the feature amount analyzing means is provided with a complexity degree calculating means, a complexity degree threshold processing means, and a complexity degree base divided region generating means. Then, the complexity of the pixel value is obtained as a feature amount for each of a plurality of predetermined regions by the complexity degree calculation means. Further, the complexity degree threshold value processing means classifies the complexity of the pixel values obtained by the complexity degree calculation means into a plurality of types using a plurality of types of threshold values. Further, the image is classified into a plurality of types based on the classification result by the complexity degree threshold processing means by the complexity degree based divided region generating means.

  For this reason, the moving image encoding apparatus of (4) can divide the image to be encoded into a plurality of types of regions using the complexity of the pixel value as the feature amount.

  (5) The present invention relates to the moving picture encoding apparatus according to (2), wherein the feature amount analyzing unit performs the feature for each of the plurality of predetermined regions (for example, equivalent to M described later). Edge direction calculation means for obtaining the edge direction as a quantity (for example, equivalent to the edge direction calculation unit 57 in FIG. 6) and the edge direction obtained by the edge direction calculation means are represented by a plurality of threshold values (for example, (N -1) edge direction threshold processing means (e.g., equivalent to the edge direction threshold processing unit 58 in FIG. 6) that classifies into a plurality of types (e.g., equivalent to N types described later) using Edge direction base divided region generating means (for example, edge direction base divided in FIG. 6) for classifying the image into a plurality of types (for example, equivalent to N types described later) based on the classification result by the edge direction threshold processing means. Have proposed that corresponds to the frequency generator 59), the moving picture coding apparatus comprising: a.

  According to the present invention, the feature quantity analyzing means includes an edge direction calculating means, an edge direction threshold processing means, and an edge direction base divided region generating means. Then, the edge direction is calculated as the feature amount for each of a plurality of predetermined regions by the edge direction calculating means. Further, the edge direction threshold processing means classifies the edge direction obtained by the edge direction calculation means into a plurality of types using a plurality of types of threshold values. Further, the edge direction base divided region generating means classifies the images into a plurality of types based on the classification result by the edge direction threshold processing means.

  For this reason, the moving image encoding apparatus of (5) can divide the image to be encoded into a plurality of types of regions using the edge direction as the feature quantity.

  (6) In the moving image encoding device according to (3) or (4), the present invention is characterized in that the feature amount analyzing unit is configured to each of the predetermined plurality of areas (for example, equivalent to M described later). In addition, an edge direction calculation unit (for example, equivalent to the edge direction calculation unit 57 in FIG. 6) that obtains an edge direction as the feature amount, and an edge direction obtained by the edge direction calculation unit are represented by a plurality of threshold values (for example, Edge direction threshold value processing means (e.g., equivalent to the edge direction threshold value processing unit 58 in FIG. 6) that classifies into a plurality of types (e.g., equivalent to N types described below) using (N-1) threshold values described later). ) And edge direction base divided region generating means for classifying the image into a plurality of types (for example, equivalent to N types described later) based on the classification result by the edge direction threshold processing means (for example, the edge direction in FIG. 6). And corresponds to the base divided area generation unit 59), it has proposed a moving picture coding apparatus further comprising a.

  According to the present invention, the feature amount analyzing means is further provided with an edge direction calculating means, an edge direction threshold processing means, and an edge direction base divided region generating means. Then, the edge direction is calculated as the feature amount for each of a plurality of predetermined regions by the edge direction calculating means. Further, the edge direction threshold processing means classifies the edge direction obtained by the edge direction calculation means into a plurality of types using a plurality of types of threshold values. Further, the edge direction base divided region generating means classifies the images into a plurality of types based on the classification result by the edge direction threshold processing means.

  For this reason, the moving picture encoding apparatus of (6) can use the edge strength and the edge direction in combination as the feature amount, or can use the complexity of the pixel value and the edge direction in combination. Therefore, compared with the case where any one of the edge strength, the complexity of the pixel value, and the edge direction is used as the feature amount, the image is divided into regions more suitable for the design, and the adaptive filter coefficient more suitable for the design Can be obtained for each divided region.

  (7) In the moving image encoding device according to any one of (1) to (6), the encoded data generation unit uses the adaptive filter processing unit in the frame as a set value for each frame. Adaptive filter coefficients (for example, corresponding to N types of adaptive filter coefficients j in FIG. 1) and the number of types of adaptive filter coefficients used by the adaptive filter processing means in the frame (for example, corresponding to N types described later) ) And the filter application information in the frame, a moving image encoding apparatus is proposed that generates encoded data having an encoded data structure.

  According to the present invention, the encoded data generating means uses, as setting values for each frame, the adaptive filter coefficients used in the adaptive filter processing means in the frame and the types of adaptive filter coefficients used in the adaptive filter processing means in the frame. The encoded data having the encoded data structure including the number and the filter application information in the frame is generated.

  For this reason, the moving image encoding apparatus according to (7) includes encoded data including the applied adaptive filter coefficients, the number of types of applied adaptive filter coefficients, and filter application information as setting values for each frame. Structured encoded data can be generated.

  (8) In the moving image encoding device according to any one of (1) to (7), the encoded data generation unit may use the adaptive filter in each of the plurality of types of regions as the filter application information. Generating the encoded data using identification information (for example, corresponding to an index number of an adaptive filter coefficient used for filter processing in each region described later) representing only the adaptive filter coefficient used by the processing means; A featured video encoding device has been proposed.

  According to the present invention, the encoded data generating means generates encoded data using the identification information representing only the adaptive filter coefficients used by the adaptive filter processing means in each of a plurality of types of regions as the filter application information. It was decided to.

  For this reason, the moving image encoding device of (8) can include identification information representing only adaptive filter coefficients applied to each of a plurality of types of regions in encoded data.

  (9) The present invention provides the moving image encoding device according to any one of (1) to (7), wherein the encoded data generation means includes all the filter coefficient calculation means obtained as the filter application information. Appropriate flag information indicating whether or not each of the plurality of types of regions has been used by the adaptive filter processing means for each of the adaptive filter coefficients (for example, N indicating the suitability of each of N types of adaptive filter coefficients j described later) In other words, a moving picture coding apparatus is proposed that generates the coded data using a corresponding flag information.

  According to the present invention, the encoded data generation means is used by the adaptive filter processing means in each of a plurality of types of areas for each of all the adaptive filter coefficients obtained by the filter coefficient calculation means as filter application information. The encoded data is generated using the propriety flag information indicating whether or not.

  For this reason, the moving image encoding apparatus of (9) indicates whether or not the encoded data is applied to each of a plurality of types of regions for each of all the adaptive filter coefficients obtained by the filter coefficient calculating means. Suitability flag information can be included.

  (10) The present invention relates to a moving picture decoding apparatus that decodes encoded data (e.g., corresponding to the original picture a in FIG. 1) generated by the moving picture encoding apparatus according to any one of (1) to (9). Decoding means (for example, equivalent to the entropy decoding unit 110 in FIG. 4) that decodes the encoded data and obtains all adaptive filter coefficients obtained by the filter coefficient calculating means and the filter application information; Filter processing means for performing filter processing on a decoded image (for example, equivalent to the pre-filter decoded image E in FIG. 4) based on all the adaptive filter coefficients obtained by the decoding means and the filter application information ( For example, a moving picture decoding apparatus is proposed, which is characterized in that it corresponds to the filter processing unit 112 in FIG.

  According to the present invention, the moving picture decoding apparatus is provided with the decoding means and the filter processing means. Then, the encoded data is decoded by the decoding means, and all the adaptive filter coefficients obtained by the filter coefficient calculating means and the filter application information are obtained. Further, the filter processing means performs the filter processing on the decoded image based on all the adaptive filter coefficients and the filter application information obtained by the decoding means.

  For this reason, the moving image decoding apparatus of (10) can perform the filtering process on the decoded image using the same adaptive filter coefficient for each area subjected to the filtering process by the adaptive filter processing unit.

  (11) The present invention is a moving image encoding method that permits adaptive filter processing and controls filter processing for each filter processing unit block in encoding processing, and is an image to be encoded (for example, the original image in FIG. 1). A first step (for example, corresponding to steps S1 to S4 in FIG. 3) that divides the image into a plurality of types (for example, corresponding to N types described later) based on the feature amount of the image a) In each of the plurality of types of regions, a second step (for example, corresponding to step S5 in FIG. 3) for obtaining an adaptive filter coefficient, and each of the plurality of types of regions in the second step A third step (for example, the step of FIG. 3) for performing a filtering process on the locally decoded image (for example, the local decoded image i of FIG. 1) using the obtained adaptive filter coefficient 6), all the adaptive filter coefficients obtained in the second step (for example, equivalent to N types of adaptive filter coefficients j in FIG. 1), and the third step among all the adaptive filter coefficients. And 4th to generate encoded data (for example, equivalent to encoded data c in FIG. 1) including filter application information (for example, equivalent to filter application information to be described later) used for each region. And a step (e.g., corresponding to a procedure for generating encoded data c by the entropy encoding unit 4 in FIG. 1).

  According to the present invention, first, based on the feature amount of the image to be encoded, the image is divided into a plurality of types of regions, and adaptive filter coefficients are obtained for each of the plurality of types of regions. Next, a filter process is performed on the locally decoded image using the obtained adaptive filter coefficient for each of a plurality of types of regions. Next, encoded data including all the adaptive filter coefficients obtained as described above and filter application information indicating, for each region, the one used in the above-described filter processing among all the adaptive filter coefficients is generated. According to this, an effect similar to the effect mentioned above can be produced.

  (12) The present invention is a program for causing a computer to execute a moving image encoding method that allows adaptive filter processing and controls filter processing for each filter processing unit block in encoding processing, A first step (for example, FIG. 3) that divides the image into a plurality of types (for example, N types to be described later) based on the feature amount of the image to be performed (for example, the original image a in FIG. 1). For each of the plurality of types of regions), a second step for obtaining an adaptive filter coefficient (for example, corresponding to step S5 in FIG. 3), and each of the plurality of types of regions. Each time, a filtering process is performed on a locally decoded image (for example, the local decoded image i in FIG. 1) using the adaptive filter coefficient obtained in the second step. A third step to be performed (for example, corresponding to step S6 in FIG. 3), all the adaptive filter coefficients obtained in the second step (for example, corresponding to N types of adaptive filter coefficients j in FIG. 1), and Encoded data (for example, the code shown in FIG. 1) including filter application information (for example, equivalent to filter application information described later) indicating, for each region, all adaptive filter coefficients used in the third step. A program for causing a computer to execute a fourth step (equivalent to a procedure for generating encoded data c by the entropy encoding unit 4 in FIG. 1), Yes.

  According to the present invention, by causing a computer to execute a program, first, an image is divided into a plurality of types of regions based on the feature amount of the image to be encoded, and an adaptive filter is applied to each of the plurality of types of regions. Find the coefficient. Next, a filter process is performed on the locally decoded image using the obtained adaptive filter coefficient for each of a plurality of types of regions. Next, encoded data including all the adaptive filter coefficients obtained as described above and filter application information indicating, for each region, the one used in the above-described filter processing among all the adaptive filter coefficients is generated. According to this, an effect similar to the effect mentioned above can be produced.

  ADVANTAGE OF THE INVENTION According to this invention, the adaptive filter coefficient which can follow a design, suppressing the increase in a processing amount can be calculated | required, and compression encoding performance can be improved.

It is a block diagram which shows the structure of the moving image encoder which concerns on 1st Embodiment of this invention. It is a block diagram which shows the structure of the feature-value analysis part with which the said moving image encoder is provided. It is a flowchart which shows operation | movement of the said moving image encoder. It is a block diagram which shows the structure of the moving image decoding apparatus which concerns on 1st Embodiment of this invention. It is a block diagram which shows the structure of the feature-value analysis part with which the moving image encoder which concerns on 2nd Embodiment of this invention is provided. It is a block diagram which shows the structure of the feature-value analysis part with which the moving image encoder which concerns on 3rd Embodiment of this invention is provided.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that the constituent elements in the following embodiments can be appropriately replaced with existing constituent elements and the like, and various variations including combinations with other existing constituent elements are possible. Accordingly, the description of the following embodiments does not limit the contents of the invention described in the claims.

<First Embodiment>
[Configuration of Moving Image Encoding Device AA]
FIG. 1 is a block diagram showing a configuration of a video encoding apparatus AA according to the first embodiment of the present invention. The moving image coding apparatus AA includes a prediction value generation unit 1, a DCT / quantization unit 2, an inverse DCT / inverse quantization unit 3, an entropy coding unit 4, a feature amount analysis unit 5, a filter coefficient calculation unit 6, an adaptive filter A processing unit 7 and a local memory 8 are provided, and encoding is performed in encoding control units (macroblock units).

  The predicted value generation unit 1 receives the original image a and the encoded image d supplied from the local memory 8 as inputs. The predicted value generation unit 1 generates and outputs a predicted value e based on a prediction method in which the highest encoding performance is expected. As a prediction method, for example, intra prediction or inter prediction can be used.

  The DCT / quantization unit 2 receives the difference between the original image a and the predicted value e as input. The DCT / quantization unit 2 performs DCT processing and quantization processing on the input signal, and outputs the result as a quantized DCT coefficient (residual signal) f.

  The inverse DCT / inverse quantization unit 3 receives the quantized DCT coefficient (residual signal) f as an input. The inverse DCT / inverse quantization unit 3 performs an inverse quantization process and an inverse DCT process on an input signal, and outputs it as a pixel signal g subjected to inverse DCT.

  The feature amount analysis unit 5 receives the original image a and the control parameter b. The control parameter b includes the area division number N (N is an integer satisfying N ≧ 2) and (N−1) types of thresholds. The feature amount analysis unit 5 has a configuration described later with reference to FIG. 2, and divides the original image a into N types of regions. Then, region division information h configured to include the region division number N and region information including information such as the shapes of the N types of regions is output.

  FIG. 2 is a block diagram illustrating a configuration of the feature amount analysis unit 5. The feature quantity analysis unit 5 includes an edge strength calculation unit 51, an edge strength threshold processing unit 52, and an edge strength base divided region generation unit 53.

  The edge strength calculation unit 51 divides the original image a into M regions (M is an integer satisfying M ≧ N), and calculates the edge strength as a feature amount for the pattern for each of the M regions. Specifically, in each of the M regions, the edge strength for each pixel in the region is calculated, and the root mean square value of the calculated edge strength is calculated as the edge strength in this region. As a method for obtaining the edge strength for each pixel, for example, a method using a Sobel filter or a Laplacian filter can be applied.

  The edge strength threshold processing unit 52 classifies the edge strengths in each of the M regions calculated by the edge strength calculation unit 51 into N types using (N−1) types of thresholds.

  Based on the classification result by the edge strength threshold processing unit 52, the edge strength base divided region generation unit 53 classifies the regions corresponding to the edge strength classified into the same type as the same type region. According to this, based on the classification result by the edge intensity threshold processing unit 52, the original image a is divided into N types of regions.

  Returning to FIG. 1, the filter coefficient calculation unit 6 receives the original image a, the area division information h, and the local decoded image i. Here, the local decoded image i is a sum of the predicted value e and the inverse DCT pixel signal g, and is a locally decoded image. The filter coefficient calculation unit 6 calculates an adaptive filter coefficient that minimizes the square error between the original image a and the local decoded image i in each region for each of the N types of regions according to the region division information h. , N types of adaptive filter coefficients j are output.

  The adaptive filter processing unit 7 receives the region division information h, the local decoded image i, and N types of adaptive filter coefficients j. The adaptive filter processing unit 7 selects one of N types of adaptive filter coefficients j for each filter processing unit block. According to this, one of N types of adaptive filter coefficients j is assigned to each of the N types of regions into which the original image a is divided. Then, the adaptive filter processing unit 7 performs filter processing on the local decoded image i using one assigned adaptive filter coefficient for each of the N types of regions. Then, the filtered local decoded image, that is, the encoded image d, and the suitability information k regarding the N types of adaptive filter coefficients j are output. The suitability information k is information on the adaptive filter coefficient used when performing the filtering process in each of the N types of regions.

  The local memory 8 receives the encoded image d. The local memory 8 stores the encoded image d and supplies it to the predicted value generation unit 1 as appropriate.

  The entropy encoding unit 4 receives the quantized DCT coefficient (residual signal) f, N types of adaptive filter coefficients j, and suitability information k. The entropy encoding unit 4 performs entropy encoding on the input signal and outputs the encoded data c. The encoded data c includes, for each frame, N types of adaptive filter coefficients j used for filter processing in this frame, the number N of types of adaptive filter coefficients j used for frame processing in this frame, Information is described in an encoded data structure including filter application information in a frame. Here, the filter application information is information indicating the adaptive filter coefficient used for the filter processing in each of the N types of regions. In the present embodiment, in particular, the filter processing is performed in each of the N types of regions. This is information that represents only the adaptive filter coefficient used in the above, and indicates the index number of the adaptive filter coefficient used for the filter processing in each region.

[Operation of Moving Image Encoding Device AA]
FIG. 3 is a flowchart showing operations of the feature amount analysis unit 5, the filter coefficient calculation unit 6, and the adaptive filter processing unit 7. Note that the processing of steps S1 to S4 described later is performed during the in-loop filter processing.

  In step S1, the moving image encoding device AA divides the original image a into M regions by the feature amount analysis unit 5, and moves the process to step S2.

  In step S2, the moving image coding apparatus AA calculates edge strength for each of the M regions by the feature amount analysis unit 5, and moves the process to step S3.

  In step S3, the moving image encoding device AA uses the feature amount analysis unit 5 to classify the edge strengths in each of the M regions into N types using (N−1) types of thresholds, and in step S4. Move processing.

  In step S4, the moving image encoding device AA uses the feature amount analysis unit 5 to set regions corresponding to the edge strengths classified into the same type based on the edge strength classification result in step S3 as regions of the same type. Classify and move to step S5. According to this, the original image a is divided into N types of regions based on the edge strength classification result in step S3.

  In step S5, the moving image encoding device AA calculates adaptive filter coefficients for each of the N types of regions by the filter coefficient calculation unit 6, and moves the process to step S6.

  In step S6, the moving image encoding device AA performs the filter process using the adaptive filter coefficient calculated in step S5 for each of the N types of regions by the adaptive filter processing unit 7, and the feature amount analyzing unit 5, The processing by the filter coefficient calculation unit 6 and the adaptive filter processing unit 7 ends.

[Configuration of Video Decoding Device BB]
FIG. 4 is a block diagram showing the configuration of the video decoding device BB according to the first embodiment of the present invention. The video decoding device BB includes an entropy decoding unit 110, a predicted value generation unit 111, a filter processing unit 112, and a memory 113, and decodes encoded data c generated in the video encoding device AA.

  The entropy decoding unit 110 receives the encoded data c. The entropy decoding unit 110 analyzes the content described in the encoded data c according to the encoded data structure, and performs entropy decoding. Then, prediction information B and residual signal C obtained as a result of entropy decoding are output.

  The predicted value generation unit 111 receives the filtered decoded image A supplied from the memory 113 and the prediction information B as inputs. The prediction value generation unit 111 determines a prediction method based on the prediction information B, generates a prediction value D using the decoded image A that has been filtered by the determined prediction method, and outputs the prediction value D. .

  The filter processing unit 112 receives the encoded data c and the pre-filter decoded image E as inputs. Here, the pre-filter decoded image E is a sum of the residual signal C and the predicted value D. The filter processing unit 112 performs a filtering process on the pre-filtered decoded image E according to the adaptive filter coefficient and the suitability information included in the encoded data c, and outputs a decoded image A that has been subjected to the filtering process.

  The memory 113 receives the decoded image A that has been filtered. The memory 113 accumulates the filtered decoded image A, and supplies it to the predicted value generation unit 111 as appropriate when performing decoding processing on undecoded blocks.

  According to the above moving image coding apparatus AA, based on the edge strength of the original image a, the original image a is divided into N types of regions, and adaptive filter coefficients are obtained for each region. For this reason, the original image a can be divided according to the design, and adaptive filter coefficients suitable for the design can be obtained for each divided region. Therefore, an adaptive filter coefficient that can follow the pattern while suppressing an increase in the processing amount can be obtained, and the compression coding performance can be improved. When the above-mentioned area division number N is set to “2” for an HD (High Definition) resolution video, The code amount can be reduced by about 15% with respect to H.264, and the code amount can be reduced by about 2% compared to the case where the technique disclosed in Non-Patent Document 2 described above is performed.

  In addition, for each frame, the moving picture coding apparatus AA includes N types of adaptive filter coefficients j used for the filter processing in this frame, and the number N of types of adaptive filter coefficients j used for the frame processing in this frame. The encoded data c having the encoded data structure including the filter application information in this frame can be generated.

  Note that the moving image encoding device AA describes the index number of the adaptive filter coefficient used for the filter processing in each of the N types of regions as the filter application information in the encoded data c. For this reason, in the encoded data c, information representing only adaptive filter coefficients used for the filter processing in each of the N types of regions is described, and information representing adaptive filter coefficients not being used is not described. Become.

  According to the above moving image decoding apparatus BB, the same adaptive filter coefficient is used for each region subjected to the filter processing by the adaptive filter processing unit 7, that is, for each of N types of regions, to the pre-filter decoded image E. Filter processing can be performed on the filter.

Second Embodiment
[Configuration of Video Encoding Device CC]
The video encoding device CC according to the second embodiment of the present invention will be described below. The moving image encoding device CC is different from the moving image encoding device AA according to the first embodiment of the present invention shown in FIG. 1 in that a feature amount analyzing unit 5A is provided instead of the feature amount analyzing unit 5. In the video encoding device CC, the same components as those of the video encoding device AA are denoted by the same reference numerals, and the description thereof is omitted.

  FIG. 5 is a block diagram illustrating a configuration of the feature amount analysis unit 5A. The feature amount analysis unit 5A includes a complexity degree calculation unit 54, a complexity degree threshold processing unit 55, and a complexity degree base divided region generation unit 56.

  The degree-of-complexity calculation unit 54 divides the original image a into M areas, and calculates the complexity of the pixel value as a feature amount related to the pattern for each of the M areas. Specifically, in each of the M areas, the complexity of the pixel values in the area is calculated, and the quantitative pixel value complexity in this area is obtained. As a method for evaluating the complexity of the pixel value, for example, a method using a dispersion value of pixel values or an activity can be applied.

  The complexity degree threshold processing unit 55 classifies the complexity of the pixel value in each of the M areas calculated by the complexity degree calculation unit 54 into N types using (N−1) types of thresholds.

  Based on the classification result by the complexity degree threshold processing unit 55, the complexity degree-based divided area generation unit 56 classifies areas corresponding to the complexity of pixel values classified into the same type as areas of the same type. According to this, the original image a is divided into N types of regions based on the classification result by the complexity degree threshold processing unit 55.

  The encoded data c generated by the above moving image encoding device CC can be decoded by the moving image decoding device BB according to the first embodiment of the present invention shown in FIG.

  According to the above moving picture coding device CC, the original image a is divided into N types of regions based on the complexity of the pixel values of the original image a, and adaptive filter coefficients are obtained for each region. For this reason, the effect similar to the effect which the moving image encoder AA can show | play can be show | played.

<Third Embodiment>
[Configuration of Moving Image Encoding Device DD]
A video encoding device DD according to the third embodiment of the present invention will be described below. The moving image encoding device DD is different from the moving image encoding device AA according to the first embodiment of the present invention shown in FIG. 1 in that a feature amount analyzing unit 5B is provided instead of the feature amount analyzing unit 5. Note that in the video encoding device DD, the same components as those of the video encoding device AA are denoted by the same reference numerals, and the description thereof is omitted.

  FIG. 6 is a block diagram illustrating a configuration of the feature amount analysis unit 5B. The feature amount analysis unit 5B includes an edge direction calculation unit 57, an edge direction threshold processing unit 58, and an edge direction base divided region generation unit 59.

  The edge direction calculation unit 57 divides the original image a into M areas, and calculates an edge direction as a feature amount related to a pattern for each of the M areas.

  Specifically, first, in each of the M regions, the edge strength for each pixel in the region is calculated in the horizontal direction, the vertical direction, and the direction at an arbitrary angle with respect to the horizontal direction or the vertical direction. As a method for calculating the edge strength for each pixel in the region in a specific direction, for example, a method using a Sobel filter can be applied. Next, for each pixel in the M regions, a direction in which the absolute value of the calculated edge intensity is maximum is obtained, and the obtained direction is set as the edge direction of the pixel.

  Next, using the edge direction of each pixel, the edge directions of the M areas are obtained by, for example, the following first method and second method. In the first method, the edge directions of each pixel in the region are totaled, and the edge direction having the largest number of totals is set as the edge direction of this region. In the second method, the edge strength of each pixel in the region is summed for each edge direction, and the edge direction having the largest total value is set as the edge direction of this region.

  The edge direction threshold value processing unit 58 classifies each edge direction of the M areas calculated by the edge direction calculation unit 57 into N types using (N−1) types of threshold values.

  Based on the classification result by the edge direction threshold processing unit 58, the edge direction base divided region generation unit 59 classifies the regions corresponding to the edge direction classified into the same type as the same type region. According to this, based on the classification result by the edge direction threshold value processing unit 58, the original image a is divided into N types of regions.

  The encoded data c generated by the above moving image encoding device DD can be decoded by the moving image decoding device BB according to the first embodiment of the present invention shown in FIG.

  According to the above moving picture coding apparatus DD, based on the edge direction of the original image a, the original image a is divided into N types of regions, and adaptive filter coefficients are obtained for each region. For this reason, the effect similar to the effect which the moving image encoder AA can show | play can be show | played.

<Fourth embodiment>
[Configuration of Video Encoding Device EE]
A video encoding device EE according to the fourth embodiment of the present invention will be described below. The moving image encoder EE differs from the moving image encoder AA according to the first embodiment of the present invention shown in FIG. 1 in that it includes a feature amount analyzer 5B in addition to the feature amount analyzer 5. In the video encoding device EE, the same components as those of the video encoding device AA are denoted by the same reference numerals, and the description thereof is omitted.

  The moving image encoding device EE includes both a feature amount analysis unit 5 and a feature amount analysis unit 5B. Therefore, the original image a is divided into P types (P is an integer satisfying P ≧ 2) by the feature amount analysis unit 5 and Q types (Q is an integer satisfying Q ≧ 2) by the feature amount analysis unit 5B. ). According to this, the original image a is divided into P × Q types of regions.

  The encoded data c generated by the above moving image encoding device EE can be decoded by the moving image decoding device BB according to the first embodiment of the present invention shown in FIG.

  According to the above moving image encoding apparatus EE, based on the edge strength and edge direction of the original image a, the original image a is divided into P × Q types of regions, and adaptive filter coefficients are obtained for each region. For this reason, the effect similar to the effect which the moving image encoder AA can show | play can be show | played.

  Further, according to the moving image encoding device EE, the original image a is divided using not only the edge strength of the original image a but also the edge direction of the original image a. For this reason, compared with the moving image coding apparatus AA that divides the original image a using only the edge strength of the original image a, the adaptive filter coefficient more suitable for the design is obtained by dividing the original image a into a region more suitable for the design. Can be obtained for each divided region. Therefore, there are cases where the code amount can be further reduced as compared with the moving picture coding apparatus AA. Specifically, when the above-described P and Q are set to “2” for an HD (High Definition) resolution video, the method shown in Non-Patent Document 2 described above is performed. On the other hand, a code amount of about 3% may be reduced.

  The processing of the moving image encoding devices AA, CC, DD, and EE of the present invention and the processing of the moving image decoding device BB are stored in a computer-readable recording medium, and the program recorded on the recording medium is stored as a moving image. The present invention can be realized by causing the encoding devices AA, CC, DD, EE, and the video decoding device BB to read and execute them.

  The above-described program is transmitted from the moving image encoding devices AA, CC, DD, EE and the moving image decoding device BB storing the program in a storage device or the like via a transmission medium or in a transmission medium. May be transmitted to other computer systems. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.

  Further, the above-described program may be for realizing a part of the above-described function. Further, even a so-called difference file (difference program) that can realize the above-described functions in combination with a program already recorded in the moving image encoding devices AA, CC, DD, EE and the moving image decoding device BB. Good.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes a design that does not depart from the gist of the present invention.

  For example, in the above-described fourth embodiment, the case where both the feature amount analysis unit 5 and the feature amount analysis unit 5B are provided in the moving image encoding apparatus has been described. However, the present invention is not limited to this, and the feature amount analysis unit 5A and the feature amount are provided. Both of the quantity analysis units 5B may be provided in the moving picture encoding device.

  In each of the above-described embodiments, the adaptive filter processing unit 7 selects one of N types of adaptive filter coefficients j for each filter processing unit block. However, the present invention is not limited to this. For example, the adequacy of each of the N types of adaptive filter coefficients j may be controlled for each filter processing unit block. According to this, a plurality of adaptive filter coefficients can be assigned to each of the N types of regions that divide the original image a, and the combinations of adaptive filter coefficients that can be assigned are for each of the N types of regions. There are 2 to the Nth power. In addition, the filter application information described in the encoded data c is not information that describes the index number of the adaptive filter coefficient, but N pieces of appropriateness flag information that indicates the appropriateness of each of the N types of adaptive filter coefficients j. Will be described. The adequacy flag information is N pieces of flag information for each of the N types of regions, and is used in the filtering process in each region for each of the N types of adaptive filter coefficients j. It is the flag information that represents. For this reason, in the encoded data c, not only the adaptive filter coefficients used for the filter processing in each of the N types of regions, but also information representing adaptive filter coefficients that are not used is described.

  Further, in each of the above-described embodiments, various information is described in the encoded data c by the entropy encoding unit 4 for each frame. However, the present invention is not limited to this. For example, it is described for each slice. Also good.

  Further, in each of the above-described embodiments, the area division number N and the above-described M, P, and Q may each be a fixed value, or may be a variable value set for each frame or each slice. .

AA, CC, DD, EE ... moving picture coding device BB ... moving picture decoding device 1 ... predicted value generation unit 2 ... DCT / quantization unit 3 ... inverse DCT / inverse quantization Unit 4... Entropy encoding unit 5, 5A, 5B... Feature amount analysis unit 51... Edge strength calculation unit 52... Edge strength threshold processing unit 53. ... complexity degree calculation unit 55 ... complexity degree threshold processing unit 56 ... complexity degree base divided region generation unit 57 ... edge direction calculation unit 58 ... edge direction threshold processing unit 59 ... edge direction Base division region generation unit 6 ... filter coefficient calculation unit 7 ... adaptive filter processing unit 8 ... local memory 110 ... entropy decoding unit 111 ... predicted value generation unit 112 ... filter processing unit 113・ ・·memory

Claims (9)

In the encoding process, a moving image encoding apparatus that permits adaptive filter processing and controls filter processing for each filter processing unit block,
A feature amount analyzing means for dividing the image into a plurality of types of regions based on the feature amount of the image to be encoded;
Filter coefficient calculating means for obtaining an adaptive filter coefficient for each of the plurality of types of regions;
Adaptive filter processing means for performing filter processing on a local decoded image using the adaptive filter coefficient obtained by the filter coefficient calculation means for each of the plurality of types of regions;
Encoded data including all adaptive filter coefficients obtained by the filter coefficient calculation means, and filter application information indicating, for each of the areas, all adaptive filter coefficients used by the adaptive filter processing means. Encoded data generation means for generating
The feature amount analyzing means includes:
Edge direction calculating means for obtaining an edge direction as the feature amount for each of a plurality of predetermined regions;
Edge direction threshold processing means for classifying the edge direction obtained by the edge direction calculating means into a plurality of types using a plurality of types of thresholds;
An edge direction base divided region generating means for classifying the image into a plurality of types based on a classification result by the edge direction threshold processing means,
A moving image encoding apparatus, wherein the image is divided into a plurality of types based on a classification result by the edge direction base divided region generation means. The moving image encoding device according to claim 1,
The feature amount analyzing means includes:
Edge strength calculating means for obtaining edge strength as the feature amount for each of the plurality of predetermined regions;
Edge strength threshold processing means for classifying the edge strength obtained by the edge strength calculation means into a plurality of types using a plurality of types of thresholds;
A moving picture coding apparatus comprising: an edge strength base divided region generating unit that classifies the image into a plurality of types based on a classification result by the edge strength threshold processing unit. The moving image encoding device according to claim 1,
The feature amount analyzing means includes:
A degree-of-complexity calculating means for determining the complexity of a pixel value as the feature amount for each of the plurality of predetermined regions;
Complexity threshold processing means for classifying the complexity of the pixel value obtained by the complexity calculating means into a plurality of types using a plurality of types of thresholds;
A moving picture coding apparatus comprising: a complexity degree base divided region generating means for classifying the image into a plurality of types based on a classification result by the complexity degree threshold processing means. In the moving image encoder according to any one of claims 1 to 3,
The encoded data generating means uses, as setting values for each frame, the adaptive filter coefficients used in the adaptive filter processing means in the frame and the number of types of adaptive filter coefficients used in the adaptive filter processing means in the frame. And encoded data having an encoded data structure including the filter application information in the frame. In the moving image encoding device according to any one of claims 1 to 4,
The encoded data generation means generates the encoded data using identification information representing only adaptive filter coefficients used by the adaptive filter processing means in each of the plurality of types of regions as the filter application information. A moving picture coding apparatus characterized by the above. In the moving image encoding device according to any one of claims 1 to 4,
Whether the encoded data generation means is used as the filter application information by the adaptive filter processing means in each of the plurality of types of areas for each of all the adaptive filter coefficients obtained by the filter coefficient calculation means. A moving picture encoding apparatus, wherein the encoded data is generated using propriety flag information representing the above. A video system comprising a video encoding device and a video decoding device,
The moving image encoding device is:
A feature amount analyzing means for dividing the image into a plurality of types of regions based on the feature amount of the image to be encoded;
Filter coefficient calculating means for obtaining an adaptive filter coefficient for each of the plurality of types of regions;
Adaptive filter processing means for performing filter processing on a local decoded image using the adaptive filter coefficient obtained by the filter coefficient calculation means for each of the plurality of types of regions;
Encoded data including all adaptive filter coefficients obtained by the filter coefficient calculation means, and filter application information indicating, for each of the areas, all adaptive filter coefficients used by the adaptive filter processing means. Encoded data generation means for generating
The feature amount analyzing means includes:
Edge direction calculating means for obtaining an edge direction as the feature amount for each of a plurality of predetermined regions;
Edge direction threshold processing means for classifying the edge direction obtained by the edge direction calculating means into a plurality of types using a plurality of types of thresholds;
An edge direction base divided region generating means for classifying the image into a plurality of types based on a classification result by the edge direction threshold processing means,
Based on the classification result by the edge direction base divided region generation means, the image is divided into a plurality of types,
The moving picture decoding device comprises:
Decoding means for decoding the encoded data to obtain all adaptive filter coefficients obtained by the filter coefficient calculating means and the filter application information;
A moving image system comprising: filter processing means for performing filter processing on a decoded image based on all the adaptive filter coefficients and the filter application information obtained by the decoding means . In the encoding process, a moving image encoding method that allows adaptive filter processing and controls the filter processing for each filter processing unit block,
A first step of dividing the image into a plurality of types of regions based on the feature amount of the image to be encoded;
A second step of obtaining an adaptive filter coefficient for each of the plurality of types of regions;
A third step of performing filtering on the locally decoded image using the adaptive filter coefficient obtained in the second step for each of the plurality of types of regions;
Generating encoded data including all adaptive filter coefficients obtained in the second step, and filter application information indicating, for each of the areas, all adaptive filter coefficients used in the third step. And a fourth step,
In the first step,
A first procedure for obtaining an edge direction as the feature amount for each of a plurality of predetermined regions;
A second procedure for classifying the edge directions obtained by the first procedure into a plurality of types using a plurality of types of threshold values;
Performing a third procedure for classifying the image into a plurality of types based on the classification result of the second procedure,
A moving image encoding method, wherein the image is divided into a plurality of types based on a classification result obtained by the third procedure. In the encoding process, a program for causing a computer to execute a moving image encoding method that allows adaptive filter processing and controls filter processing for each filter processing unit block,
A first step of dividing the image into a plurality of types of regions based on the feature amount of the image to be encoded;
A second step of obtaining an adaptive filter coefficient for each of the plurality of types of regions;
A third step of performing filtering on the locally decoded image using the adaptive filter coefficient obtained in the second step for each of the plurality of types of regions;
Generating encoded data including all adaptive filter coefficients obtained in the second step, and filter application information indicating, for each of the areas, all adaptive filter coefficients used in the third step. And causing the computer to execute a fourth step of:
In the first step,
A first procedure for obtaining an edge direction as the feature amount for each of a plurality of predetermined regions;
A second procedure for classifying the edge directions obtained by the first procedure into a plurality of types using a plurality of types of threshold values;
Performing a third procedure for classifying the image into a plurality of types based on the classification result of the second procedure,
A program for dividing the image into a plurality of types based on the classification result according to the third procedure.

Images