WO2016129185A1

WO2016129185A1 - Moving image coding device, moving image decoding device, moving image processing system, moving image coding method, moving image decoding method, and program

Info

Publication number: WO2016129185A1
Application number: PCT/JP2015/085739
Authority: WO
Inventors: 知伸吉野
Original assignee: Kddi株式会社
Priority date: 2015-02-09
Filing date: 2015-12-22
Publication date: 2016-08-18
Also published as: JP2016146572A

Abstract

A moving image coding device 1 is provided with a prior intra prediction value calculation unit 21, an error minimization filter coefficient calculation unit 22, and a filter processing unit 23. The prior intra prediction value calculation unit 21 generates a prediction value by intra prediction based on an intra prediction mode using an input image SIG1 and a locally decoded image SIG17, and outputs the prediction value as a prior intra prediction value SIG18. The error minimization filter coefficient calculation unit 22 calculates an interpolation filter coefficient for minimizing an error based on the input image SIG1 and the prior intra prediction value SIG18, and outputs the interpolation filter coefficient as an interpolation filter coefficient SIG8. The filter processing unit 23 performs error minimization filter processing on the prior intra prediction value SIG18 using the interpolation filter coefficient SIG8.

Description

Moving picture coding apparatus, moving picture decoding apparatus, moving picture processing system, moving picture coding method, moving picture decoding method, and program

The present invention relates to a moving image encoding device, a moving image decoding device, a moving image processing system, a moving image encoding method, a moving image decoding method, and a program.

In a standard method related to video compression represented by Non-Patent Document 1, a predictive coding method is adopted, and a prediction error is quantized after being converted into a frequency domain. For this reason, quantization errors (noise) occur in the encoded image during the quantization process. Therefore, a method for directly or indirectly suppressing the coding error has been proposed.

For example, Non-Patent Document 1 proposes SAO (Sample Adaptive Offset). SAO is a type of filter that controls an offset value for each local region with respect to an encoded frame.

Also, Non-Patent Document 2 proposes ALF (Adaptive Loop Filter). In ALF, a filter process that minimizes an error for each local region is performed on an encoded frame.

SAO in Non-Patent Document 1 and ALF in Non-Patent Document 2 are techniques for encoded frames, that is, techniques for frames that have already been encoded, and strongly affect the compression performance in predictive encoding. It is not a technology that improves prediction performance. For this reason, the compression performance could not be improved sufficiently.

Therefore, the present invention has been made in view of the above-described problems, and an object thereof is to improve the compression performance.

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

(1) The present invention is a video encoding device (for example, the video code shown in FIG. 1) that allows predictive encoding using an encoded local decoded image (eg, equivalent to the local decoded image SIG17 shown in FIG. 2). Between the inter prediction based on integer-precision motion compensation and the intra prediction using the input image (e.g., equivalent to the input image SIG1 in FIG. 2) and the local decoded image. Pre-prediction value generation means (for example, equivalent to the pre-intra prediction value calculation unit 21 in FIG. 3) that generates a prediction value by at least one of the above, the input image, and the prediction value generated by the prior prediction value generation means And error minimizing filter coefficient calculating means (for example, the error minimizing filter shown in FIG. 3) for calculating a filter coefficient for minimizing an error based on Filter processing means for performing filter processing on the prediction value generated by the prior prediction value generation means using the filter coefficient calculated by the error minimization filter coefficient calculation means. (For example, corresponding to the filter processing unit 23 in FIG. 3) is proposed.

According to the present invention, the moving picture coding apparatus is provided with the pre-predicted value generating means, the error minimizing filter coefficient calculating means, and the filter processing means. In addition, the pre-prediction value generation means generates the prediction value by at least one of inter prediction based on integer precision motion compensation and intra prediction, using the input image and the local decoded image. Further, the error minimizing filter coefficient calculating means calculates the filter coefficient that minimizes the error based on the input image and the predicted value generated by the prior predicted value generating means. Further, the filter processing means performs the filter processing on the prediction value generated by the prior prediction value generation means using the filter coefficient calculated by the error minimization filter coefficient calculation means.

For this reason, the filtering process is performed on the frame before being encoded, that is, on the frame before the encoding process is performed. Therefore, it is possible to improve the prediction performance that strongly affects the compression performance in predictive coding, and as a result, it is possible to improve the compression performance.

Also, a filter coefficient that minimizes at least one of an error based on the input image and a prediction value based on inter prediction based on integer-precision motion compensation and an error based on the prediction value based on the input image and intra prediction is calculated. Will be. As a result, errors in integer pixel position motion compensation and intra prediction can be reduced, and as a result, compression performance can be further improved.

(2) In the moving picture coding apparatus according to (1), the error minimizing filter coefficient calculation unit is a difference between the input image and the prediction value generated by the prior prediction value generation unit. A filter that minimizes any one of a prediction error and an encoding error obtained by performing orthogonal transform processing, quantization processing, inverse quantization processing, and inverse transform processing on the prediction error A moving picture coding apparatus characterized by calculating coefficients is proposed.

According to the present invention, in the moving picture coding apparatus of (1), the filter coefficient for minimizing either the prediction error or the coding error is calculated by the error minimizing filter coefficient calculation means. For this reason, the prediction error or the coding error can be minimized by the filter processing by the filter processing means.

Note that when the prediction error is minimized, it is not necessary to perform orthogonal transformation processing, quantization processing, inverse quantization processing, and inverse transformation processing as compared with the case of minimizing the coding error. It is possible to reduce the amount of processing in the calculation filter coefficient calculation means.

On the other hand, when the coding error is minimized, the filter coefficient can be obtained in consideration of not only the prediction error but also the quantization error. For this reason, it is possible to simultaneously realize the suppression of the code amount of the prediction residual and the suppression of the direct error in the decoded image. Therefore, compared with the case where the prediction error is minimized, the compression performance can be particularly improved in intra coding in which a large amount of prediction residual code is generated, high bit rate coding, or the like.

(3) The present invention relates to the moving picture coding apparatus according to (1) or (2), wherein the error minimizing filter coefficient calculating means calculates the filter coefficient based on a Wiener filter method. A video encoding device is proposed.

According to the present invention, the filter coefficient can be calculated based on the Wiener filter method in the moving picture coding apparatus of (1) or (2).

(4) The present invention relates to the moving picture encoding apparatus according to any one of (1) to (3), wherein the prior prediction value generation means performs prediction by intra prediction using the input image and the local decoded image. A value is generated, and the error minimizing filter coefficient calculation means proposes a moving picture coding apparatus characterized in that the filter coefficient is calculated according to an intra prediction mode.

Where H. H.264 and H.264. In the standard method related to video compression represented by H.265, an intra prediction value corresponding to the intra prediction mode is generated in the intra prediction. For this reason, the intra prediction value is easily affected by the intra prediction mode.

Therefore, according to the present invention, in any one of the video encoding apparatuses (1) to (3), the filter coefficient is calculated according to the intra prediction mode by the prior prediction value generation means. For this reason, compared with the case where one filter coefficient common to all intra prediction modes is calculated, a prediction error can be made small or encoding performance can be improved.

(5) The present invention relates to the moving picture coding apparatus according to (4), wherein the error minimizing filter coefficient calculation means includes m intra prediction modes (where m is smaller than the number of types of intra prediction modes, A moving picture coding apparatus has been proposed in which a filter coefficient is calculated for each classified group.

Here, when the filter coefficient is calculated according to each of the intra prediction modes, the amount of information regarding the filter coefficient becomes enormous, and as a result, the encoding performance may be degraded.

Therefore, according to the present invention, in the moving picture encoding apparatus of (4), the intra prediction mode is classified into m groups by the error minimizing filter coefficient calculation means, and the filter coefficient is calculated for each classified group. It was decided. For this reason, compared with the case where one filter coefficient common to all intra prediction modes is calculated, it is possible to improve the encoding performance by suppressing an increase in the amount of information regarding the filter coefficient while reducing the prediction error. it can.

(6) The present invention is a video decoding device (for example, equivalent to the video decoding device 100 in FIG. 1) that allows predictive decoding using a decoded image (eg, equivalent to the decoded image SIG64 in FIG. 4). Entropy decoding means (for example, corresponding to the interpolation filter coefficient SIG54 in the intra prediction of FIG. 4) by entropy decoding the encoded data (for example, corresponding to the encoded data SIG2 of FIG. 4) (for example, 4 corresponding to the entropy decoding unit 110 in FIG. 4), inter prediction based on integer-precision motion compensation, and intra prediction, and generating a prediction value using the decoded image, and generating the prediction value Filter processing means (for example, performing filter processing using the filter coefficients obtained by the entropy decoding means) Proposes a moving picture decoding apparatus, characterized in that it comprises a corresponding) to the intra prediction value generation unit 140 of FIG. 4, a.

According to the present invention, the moving picture decoding apparatus is provided with entropy decoding means and filter processing means. In addition, the entropy decoding means entropy decodes the encoded data to obtain filter coefficients. Further, the filter processing unit generates a prediction value using the decoded image by at least one of inter prediction based on integer-precision motion compensation and intra prediction, and entropy decoding is performed on the generated prediction value. The filter processing is performed using the filter coefficient acquired by the means.

For this reason, the filter coefficient used when generating the encoded data is used for the prediction value generated by at least one of inter prediction based on integer precision motion compensation and intra prediction. Processing can be performed. Therefore, errors in motion compensation at integer pixel positions and intra prediction can be reduced, and as a result, compression performance can be improved.

(7) In the video decoding device according to (6), the entropy decoding unit performs entropy decoding on the encoded data, obtains a filter coefficient calculated according to an intra prediction mode, and performs the filtering process The means proposes a moving picture decoding apparatus characterized in that a predicted value is generated by intra prediction using the decoded picture.

Therefore, according to the present invention, in the moving picture decoding apparatus of (6), the encoded data is entropy decoded by the entropy decoding means, and the filter coefficient calculated according to the intra prediction mode is obtained. In addition, a prediction value is generated by intra prediction using the decoded image by the filter processing means. For this reason, since it is possible to perform the filter processing using the filter coefficient calculated according to the intra prediction mode, compared to the case of performing the filter processing using one filter coefficient common to all intra prediction modes, The prediction error can be reduced and the encoding performance can be improved.

(8) The present invention relates to the moving picture decoding apparatus according to (7), wherein the number of intra prediction modes is m in advance (where m is a natural number that is smaller than the number of types of intra prediction modes and does not include zero). The filter coefficient is calculated for each group that is classified, and the entropy decoding means performs entropy decoding on the encoded data to obtain the filter coefficient calculated for each group. A moving image decoding apparatus having a feature is proposed.

Here, when the filter coefficient is calculated according to each intra prediction mode, the amount of information regarding the filter coefficient becomes enormous, and as a result, the encoding performance may be degraded.

Therefore, according to the present invention, in the moving picture decoding apparatus of (7), the intra prediction modes are classified into m groups in advance, and the filter coefficients are calculated for each classified group, and entropy decoding is performed. The means entropy-decodes the encoded data and obtains a filter coefficient calculated for each group. For this reason, compared with the case where filter processing is performed using one filter coefficient common to all intra prediction modes, an increase in the amount of information regarding the filter coefficient is suppressed while reducing the prediction error, and the encoding performance is reduced. Can be improved.

(9) The present invention encodes a moving image to generate encoded data (for example, equivalent to the encoded data SIG2 in FIG. 2) (for example, the moving image encoding device 1 in FIG. 1). And a moving picture decoding apparatus (for example, equivalent to the moving picture decoding apparatus 100 in FIG. 1) that decodes encoded data generated by the moving picture encoding apparatus. 1 is equivalent to the moving image processing system AA), and the moving image encoding apparatus includes an input image (for example, corresponding to the input image SIG1 in FIG. 2) and a local decoded image (for example, the local decoded image SIG17 in FIG. 2). 3), a pre-prediction value generation unit (for example, the pre-intra prediction unit of FIG. 3) that generates a prediction value by at least one of inter prediction based on integer-precision motion compensation and intra prediction. An error (e.g., corresponding to a prediction error or an encoding error described later) based on the input image and the prediction value generated by the prior prediction value generation means. With respect to the prediction value generated by the error minimization filter coefficient calculation means (for example, equivalent to the error minimization filter coefficient calculation unit 22 in FIG. 3) for calculating the filter coefficient, and the prediction value generated by the prior prediction value generation means, the error An encoding-side filter processing means (for example, corresponding to the filter processing unit 23 in FIG. 3) that performs filter processing using the filter coefficient calculated by the minimizing filter coefficient calculating means, and the error minimizing filter coefficient calculating means. Entropy encoding means for generating encoded data by entropy encoding the filtered filter coefficients (for example, equivalent to the entropy encoding unit 40 in FIG. 2) And the moving picture decoding apparatus entropy-decodes the encoded data generated by the entropy encoding means and obtains the filter coefficient calculated by the error minimizing filter coefficient calculation means ( For example, it is equivalent to the entropy decoding unit 110 in FIG. 4), inter prediction based on integer-precision motion compensation, and intra prediction, which is used when the predicted value is generated by the prior predicted value generation unit. 4 generates a predicted value using the decoded image, and performs a filtering process on the generated predicted value using the filter coefficient acquired by the entropy decoding unit (for example, the intra filter in FIG. 4). And a predicted value generation unit 140).

According to the present invention, the moving image processing system is provided with a moving image encoding device and a moving image decoding device, and the moving image encoding device includes pre-predicted value generation means, error minimizing filter coefficient calculation means, and encoding side filter. Processing means and entropy encoding means are provided. In addition, the pre-prediction value generation means generates the prediction value by at least one of inter prediction based on integer precision motion compensation and intra prediction, using the input image and the local decoded image. Further, the error minimizing filter coefficient calculating means calculates the filter coefficient that minimizes the error based on the input image and the predicted value generated by the prior predicted value generating means. Further, the encoding-side filter processing means performs filter processing on the prediction value generated by the prior prediction value generation means using the filter coefficient calculated by the error minimization filter coefficient calculation means. Further, the entropy encoding unit entropy encodes the filter coefficient calculated by the error minimizing filter coefficient calculation unit to generate encoded data. The moving picture decoding apparatus is provided with entropy decoding means and decoding side filter processing means. The entropy decoding unit entropy-decodes the encoded data generated by the entropy encoding unit, and obtains the filter coefficient calculated by the error minimizing filter coefficient calculation unit. Further, the decoding-side filter processing means uses the decoded image by the one used when the prediction value is generated by the prior prediction value generation means among inter prediction based on integer-precision motion compensation and intra prediction. A prediction value is generated, and filter processing is performed on the generated prediction value using the filter coefficient acquired by the entropy decoding unit. According to this, an effect similar to the effect mentioned above can be produced.

(10) The present invention is based on prior prediction value generation means (for example, equivalent to the prior intra prediction value calculation section 21 in FIG. 3) and error minimization filter coefficient calculation means (for example, error minimization filter coefficient calculation section 22 in FIG. 3). ) And filter processing means (for example, equivalent to the filter processing unit 23 in FIG. 3), and a prediction code using an encoded local decoded image (for example, equivalent to the local decoded image SIG17 in FIG. 2) In the moving picture coding method in the moving picture coding apparatus (for example, equivalent to the moving picture coding apparatus 1 in FIG. 1) that permits the conversion, the prior prediction value generation means is configured to input the input picture (for example, FIG. 2). 1) and a local decoded image, and a first step for generating a prediction value by at least one of inter prediction based on integer-precision motion compensation and intra prediction. And the error minimizing filter coefficient calculation means minimizes an error (e.g., corresponding to a prediction error or an encoding error described later) based on the input image and the prediction value generated by the first step. A second step of calculating a filter coefficient to be converted, and the filter processing means performs a filtering process on the prediction value generated in the first step using the filter coefficient calculated in the second step And a third step of performing a video encoding method.

According to the present invention, a predicted value is generated by at least one of inter prediction based on integer precision motion compensation and intra prediction using an input image and a locally decoded image. In addition, a filter coefficient that minimizes an error based on the input image and the predicted value described above is calculated, and the filter process is performed on the predicted value using the calculated filter coefficient. According to this, an effect similar to the effect mentioned above can be produced.

(11) The present invention includes entropy decoding means (for example, equivalent to the entropy decoding unit 110 in FIG. 4) and filter processing means (for example, equivalent to the intra-predicted value generation unit 140 in FIG. 4). 4 is a moving picture decoding method in a moving picture decoding apparatus (for example, equivalent to the moving picture decoding apparatus 100 in FIG. 1) that allows predictive decoding using the decoded picture SIG64 in FIG. A first step of entropy decoding encoded data (for example, equivalent to the encoded data SIG2 in FIG. 4) to obtain filter coefficients (for example, equivalent to the interpolation filter coefficient SIG54 in the intra prediction in FIG. 4); The filter processing means is at least one of inter prediction based on integer precision motion compensation and intra prediction. A second step of generating a prediction value using the decoded image, and performing a filtering process on the generated prediction value using the filter coefficient acquired in the first step. A moving image decoding method is proposed.

According to the present invention, encoded data is entropy decoded to obtain a filter coefficient, and a prediction value is obtained using a decoded image by at least one of inter prediction based on integer-precision motion compensation and intra prediction. The generated prediction value is subjected to filter processing using the acquired filter coefficient. According to this, an effect similar to the effect mentioned above can be produced.

(12) The present invention is based on prior prediction value generation means (for example, equivalent to the prior intra prediction value calculation section 21 in FIG. 3) and error minimization filter coefficient calculation means (for example, error minimization filter coefficient calculation section 22 in FIG. 3). ) And filter processing means (for example, equivalent to the filter processing unit 23 in FIG. 3), and a prediction code using an encoded local decoded image (for example, equivalent to the local decoded image SIG17 in FIG. 2) A program for causing a computer to execute a moving picture coding method in a moving picture coding apparatus (for example, equivalent to the moving picture coding apparatus 1 in FIG. 1) that permits the encoding, the prior prediction value generation means However, using an input image (e.g., equivalent to the input image SIG1 in FIG. 2) and the local decoded image, the number of inter predictions based on integer-precision motion compensation and intra predictions is small. In either case, the first step of generating a predicted value, and the error minimizing filter coefficient calculation means calculates an error based on the input image and the predicted value generated by the first step (for example, described later). A second step of calculating a filter coefficient that minimizes a prediction error or a coding error of the second step, and the filter processing means for the predicted value generated by the first step A program for causing a computer to execute a third step of performing filter processing using the filter coefficient calculated in the step is proposed.

According to the present invention, by executing a program using a computer, a prediction value is determined by at least one of inter prediction based on integer-precision motion compensation and intra prediction using an input image and a locally decoded image. It was decided to generate. In addition, a filter coefficient that minimizes an error based on the input image and the predicted value described above is calculated, and the filter process is performed on the predicted value using the calculated filter coefficient. According to this, an effect similar to the effect mentioned above can be produced.

(13) The present invention includes an entropy decoding unit (for example, equivalent to the entropy decoding unit 110 in FIG. 4) and a filter processing unit (for example, equivalent to the intra prediction value generation unit 140 in FIG. 4), and a decoded image (for example, , A program for causing a computer to execute a moving picture decoding method in a moving picture decoding apparatus (for example, equivalent to the moving picture decoding apparatus 100 in FIG. 1) that allows predictive decoding using the decoded picture SIG64 in FIG. The entropy decoding means entropy-decodes the encoded data (for example, equivalent to the encoded data SIG2 in FIG. 4) to obtain filter coefficients (for example, equivalent to the interpolation filter coefficient SIG54 in the intra prediction in FIG. 4). A first step of obtaining and inter-prediction wherein the filtering means is based on integer precision motion compensation In addition, the prediction value is generated using the decoded image by at least one of intra prediction and intra-prediction, and the generated prediction value is subjected to filter processing using the filter coefficient acquired in the first step. A program for causing a computer to execute the second step is proposed.

According to the present invention, by executing a program using a computer, the encoded data is entropy-decoded to obtain filter coefficients, and at least one of inter prediction based on integer precision motion compensation and intra prediction is used. Thus, a predicted value is generated using the decoded image, and a filter process is performed on the generated predicted value using the acquired filter coefficient. According to this, an effect similar to the effect mentioned above can be produced.

According to the present invention, the compression performance can be improved.

1 is a block diagram of a moving image processing system according to a first embodiment of the present invention. 1 is a block diagram of a video encoding apparatus according to a first embodiment of the present invention. It is a block diagram of the intra estimated value production | generation part with which the moving image encoder which concerns on 1st Embodiment of this invention is provided. It is a block diagram of the moving image decoding apparatus which concerns on 1st Embodiment of this invention. It is a block diagram of the moving image processing system which concerns on 2nd Embodiment of this invention. It is a block diagram of the moving image encoder which concerns on 2nd Embodiment of this invention. It is a block diagram of the integer precision MC prediction value production | generation part with which the moving image encoder which concerns on 2nd Embodiment of this invention is provided. It is a block diagram of the moving image decoding apparatus which concerns on 2nd Embodiment of this invention.

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 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>
FIG. 1 is a block diagram of a moving image processing system AA according to the first embodiment of the present invention. The moving image processing system AA encodes a moving image to generate encoded data (see encoded data SIG2 in FIGS. 2 and 4), and a code generated by the moving image encoding device 1. And a moving picture decoding apparatus 100 for decoding the digitized data. These moving image encoding apparatus 1 and moving image decoding apparatus 100 transmit and receive the above-described encoded data via a transmission path, for example.

[Configuration and Operation of Video Encoding Device 1]
FIG. 2 is a block diagram of the moving picture encoding apparatus 1. The moving image encoding apparatus 1 includes an inter prediction value generation unit 10, an intra prediction value generation unit 20, a DCT / quantization unit 30, an entropy encoding unit 40, an inverse DCT / inverse quantization unit 50, and a local memory 60. .

The inter prediction value generation unit 10 includes a decimal accuracy MC prediction value generation unit 11 and an integer accuracy MC prediction value generation unit 12.

The decimal precision MC predicted value generation unit 11 receives an input image SIG1 and a later-described local decoded image SIG17 supplied from the local memory 60 as inputs. First, the decimal precision MC prediction value generation unit 11 adaptively calculates an interpolation filter coefficient that minimizes a prediction error caused by decimal precision motion compensation (MC) using the input image SIG1 and the local decoded image SIG17. In addition to outputting as an interpolation filter coefficient SIG3, control information (including a decimal precision motion vector) SIG4 relating to decimal precision inter prediction is output. As a method for calculating the interpolation filter coefficient, for example, a calculation method based on the Wiener filter method can be used. Next, the decimal precision MC prediction value generation unit 11 generates a prediction value based on the decimal precision MC using the calculated interpolation filter coefficient, and outputs the prediction value as an inter prediction value SIG5 based on the decimal precision MC.

The integer precision MC predicted value generation unit 12 receives the input image SIG1 and the locally decoded image SIG17 supplied from the local memory 60. The integer accuracy MC predicted value generation unit 12 generates a predicted value based on the integer accuracy MC using the input image SIG1 and the local decoded image SIG17, and outputs the predicted value as an inter predicted value SIG7 based on the integer accuracy MC. Control information (including an integer-precision motion vector) SIG6 is output.

The intra predicted value generation unit 20 receives the input image SIG1 and the local decoded image SIG17 supplied from the local memory 60 as inputs. The intra prediction value generation unit 20 generates a prediction value based on intra prediction using the input image SIG1 and the locally decoded image SIG17, and then performs an error minimization filter process described later to output an intra prediction value SIG9. Then, an interpolation filter coefficient SIG8, which will be described later, and control information (including an intra prediction mode) SIG10 related to intra prediction are output. The intra predicted value generation unit 20 will be described in detail below with reference to FIG.

FIG. 3 is a block diagram of the intra prediction value generation unit 20. The intra prediction value generation unit 20 includes a prior intra prediction value calculation unit 21, an error minimization filter coefficient calculation unit 22, and a filter processing unit 23.

The prior intra prediction value calculation unit 21 receives the input image SIG1 and the locally decoded image SIG17. The pre-intra prediction value calculation unit 21 generates a prediction value by intra prediction based on the intra prediction mode using the input image SIG1 and the locally decoded image SIG17, and outputs the prediction value as the pre-intra prediction value SIG18. The control information SIG10 regarding is output.

The error minimizing filter coefficient calculation unit 22 receives the input image SIG1 and the prior intra prediction value SIG18. The error minimizing filter coefficient calculation unit 22 calculates an interpolation filter coefficient that minimizes an error based on the input image SIG1 and the pre-intra prediction value SIG18, and outputs the interpolation filter coefficient SIG8.

Note that the error based on the input image SIG1 and the pre-intra prediction value SIG18 is a prediction error or a coding error.

The prediction error is a difference between the input image SIG1 and the prior intra prediction value SIG18 (an error caused by intra prediction). When calculating an interpolation filter coefficient that minimizes the prediction error, the error minimization filter coefficient calculation unit 22 calculates an interpolation filter coefficient that minimizes the difference between the input image SIG1 and the prior intra prediction value SIG18. As a method for calculating the interpolation filter coefficient, for example, a calculation method based on the Wiener filter method can be used.

On the other hand, the coding error is an error obtained by combining the above-described prediction error and the quantization error generated in the quantization process. When calculating the interpolation filter coefficient that minimizes the coding error, the error minimizing filter coefficient calculation unit 22 first obtains a difference between the input image SIG1 and the prior intra prediction value SIG18. Next, after the same processing as the DCT / quantization unit 30 is performed on the obtained difference, the same processing as that of the inverse DCT / inverse quantization unit 50 is performed to obtain the above-described encoding error. Next, an interpolation filter coefficient that minimizes the obtained coding error is calculated. As a method for calculating the interpolation filter coefficient, for example, a calculation method based on the Wiener filter method can be used.

The filter processing unit 23 receives the interpolation filter coefficient SIG8 and the prior intra predicted value SIG18 as inputs. The filter processing unit 23 performs an error minimizing filter process on the previous intra predicted value SIG18 using the interpolation filter coefficient SIG8, and outputs the predicted value after the filter process as a final intra predicted value SIG9.

In addition, as a method for calculating the interpolation filter coefficient SIG8, a method for calculating one interpolation filter coefficient common to all intra prediction modes (for example, DC prediction, Planar prediction, and 33 kinds of direction predictions), intra, There is a method of calculating the interpolation filter coefficient according to the prediction mode.

Where H. H.264 and H.264. In the standard method related to video compression represented by H.265, an intra prediction value corresponding to the intra prediction mode is generated in the intra prediction. For this reason, the intra prediction value is easily affected by the intra prediction mode. Therefore, when the interpolation filter coefficient is calculated according to each intra prediction mode, the prediction error can be reduced compared to the case where one interpolation filter coefficient common to all intra prediction modes is calculated. The effect and the effect that the encoding performance can be improved can be expected.

However, when the interpolation filter coefficient is calculated according to each intra prediction mode, the amount of information related to the interpolation filter coefficient becomes enormous, and as a result, the encoding performance may be degraded.

Therefore, the intra prediction value generation unit 20 classifies the intra prediction modes into m groups (where m is a natural number that is smaller than the number of types of intra prediction modes and does not include zero), and interpolates for each group. Calculate filter coefficients. In addition, the intra prediction value generation unit 20 uses the interpolation filter coefficient SIG8 to indicate not only the interpolation filter coefficient calculated for each group but also which group each of all intra prediction modes belongs to among the m groups. Also output.

2, the DCT / quantization unit 30 receives the prediction residual signal SIG12 as an input. The prediction residual signal SIG12 is a difference signal between the input image SIG1 and the prediction value SIG11. The prediction value SIG11 is an inter prediction value SIG5 based on decimal precision MC and inter prediction based on integer precision MC. Of the value SIG7 and the intra prediction value SIG9, it is a prediction value calculated by a prediction method that is expected to have the highest coding performance. The DCT / quantization unit 30 performs an orthogonal transform process on the prediction residual signal SIG12, performs a quantization process on the transform coefficient obtained by the orthogonal transform process, and outputs the quantized transform coefficient SIG13. To do.

The entropy encoding unit 40 receives the quantized transform coefficient SIG13, the interpolation filter coefficients SIG3 and SIG8, and the control information SIG14. The control information SIG14 is expected to have the above-described highest encoding performance among the control information SIG4 related to decimal precision inter prediction, the control information SIG6 related to integer precision inter prediction, and the control information SIG10 related to intra prediction. This is control information obtained by the prediction method. The entropy encoding unit 40 performs entropy encoding processing on the input information, describes the result in the encoded data according to the description rule (encoding syntax) for the encoded data, Output as encoded data SIG2.

The inverse DCT / inverse quantization unit 50 receives the quantized transform coefficient SIG13 as input. The inverse DCT / inverse quantization unit 50 performs an inverse quantization process on the quantized transform coefficient SIG13, performs an inverse transform process on the transform coefficient obtained by the inverse quantization process, and performs an inverse process. Output as transform coefficient SIG15 subjected to orthogonal transform.

The local memory 60 receives the locally decoded image SIG16. The local decoded image SIG16 is a signal obtained by adding the predicted value SIG11 and the transform coefficient SIG15 subjected to inverse orthogonal transform. The local memory 60 accumulates the input local decoded image SIG16, and when it is necessary to refer to the past local decoded image SIG16 after the next coding processing unit block, the decimal precision MC prediction value generation unit is appropriately selected. 11 and the integer accuracy MC prediction value generation unit 12 and the intra prediction value generation unit 20 are supplied as a locally decoded image SIG17.

[Configuration and Operation of Video Decoding Device 100]
FIG. 4 is a block diagram of the video decoding device 100. The moving picture decoding apparatus 100 includes an entropy decoding unit 110, an inverse DCT / inverse quantization unit 120, an inter prediction value generation unit 130, an intra prediction value generation unit 140, and a local memory 150.

The entropy decoding unit 110 receives the encoded data SIG2. The entropy decoding unit 110 analyzes the content described in the encoded data SIG2 according to the encoded data structure and performs entropy decoding, and performs the residual signal SIG51 and the interpolation filter coefficient SIG52 (interpolation filter coefficient SIG3 in the decimal precision MC). ), Control information SIG53, and interpolation filter coefficient SIG54 (interpolation filter coefficient SIG8) in the intra prediction are acquired and output.

The inverse DCT / inverse quantization unit 120 receives the residual signal SIG51. The inverse DCT / inverse quantization unit 120 performs an inverse quantization process on the residual signal SIG51, performs an inverse transform process on the result obtained from the inverse quantization process, and performs an inverse orthogonal transform result SIG55. Output as.

The inter prediction value generation unit 130 includes a decimal accuracy MC prediction value generation unit 131 and an integer accuracy MC prediction value generation unit 132.

The decimal precision MC predicted value generation unit 131 includes an interpolation filter coefficient SIG52 in the decimal precision MC, control information SIG56 (control information SIG4 related to decimal precision inter prediction) included in the control information SIG53, A decoded image SIG64 (described later) supplied from the local memory 150 is used as an input. The decimal accuracy MC predicted value generation unit 131 first generates a predicted value based on the decimal accuracy MC in the processing block by referring to the decoded image SIG64 based on the control information SIG56 related to the decimal prediction inter prediction. Next, the generated prediction value based on the decimal precision MC is subjected to filter processing using the interpolation filter coefficient SIG52 in the decimal precision MC, and the prediction value after the filter processing is set as an inter prediction value SIG59 based on the decimal precision MC. Output.

The integer precision MC predicted value generation unit 132 includes control information SIG57 (control information SIG6 related to integer precision inter prediction) included in the control information SIG53, and a decoded image SIG64 supplied from the local memory 150. And are input. The integer precision MC predicted value generation unit 132 refers to the decoded image SIG64 based on the control information SIG57 related to integer precision inter prediction, generates a predicted value based on the integer precision MC in the processing block, and is based on the integer precision MC. It outputs as inter prediction value SIG60.

The intra prediction value generation unit 140 includes control information SIG58 related to intra prediction (control information SIG10 related to intra prediction) included in the control information SIG53, an interpolation filter coefficient SIG54 for intra prediction, and a decoded image supplied from the local memory 150. SIG64 as an input. First, the intra prediction value generation unit 140 refers to the decoded image SIG64 based on the control information SIG58 related to intra prediction, and generates an intra prediction value in the processing block. Next, the generated intra prediction value is filtered using the interpolation filter coefficient SIG54 in the intra prediction, and the prediction value after the filter processing is output as the intra prediction value SIG61.

Note that the interpolation filter coefficient SIG54 in intra prediction includes m interpolation filter coefficients and information indicating to which of the m groups the m interpolation filter coefficients belong. It is. For this reason, the intra prediction value generation unit 140 determines which group of the m groups the intra prediction mode of the processing block belongs to, and selects the interpolation filter coefficient belonging to the determined group from the m interpolation filter coefficients. The filter processing in the processing block is performed using the selected interpolation filter coefficient.

A signal obtained by adding the inverse orthogonal transform result SIG55 and the predicted value SIG62 is output from the video decoding device 100 as a decoded image SIG63 and also supplied to the local memory 150. Note that the predicted value SIG62 is a predicted method that is expected to have the highest coding performance among the inter predicted value SIG59 based on the decimal precision MC, the inter predicted value SIG60 based on the integer precision MC, and the intra predicted value SIG61. Is the predicted value calculated by

The local memory 150 receives the decoded image SIG63. The local memory 150 stores the input decoded image SIG63, and when it is necessary to refer to the past decoded image SIG63 after the next decoding processing unit block, the decimal precision MC predicted value generation unit 131 and the integer are appropriately selected. The decoded image SIG64 is supplied to the accuracy MC prediction value generation unit 132 and the intra prediction value generation unit 140.

According to the above video encoding device 1 and video decoding device 100, the following effects can be obtained.

The moving image encoding apparatus 1 performs filtering processing on the frame before encoding, that is, the frame before encoding processing is performed by the filter processing unit 23. For this reason, the prediction performance that strongly affects the compression performance in predictive coding can be improved, and as a result, the compression performance can be improved.

Further, the moving image encoding apparatus 1 calculates an interpolation filter coefficient that minimizes an error based on the input image SIG1 and the pre-intra prediction value SIG18 by the error minimizing filter coefficient calculation unit 22, and the filter processing unit 23 calculates the error Filter processing is performed using the interpolation filter coefficient calculated by the minimized filter coefficient calculation unit 22. In the moving image decoding apparatus 100, the intra prediction value generation unit 140 performs filter processing using the interpolation filter coefficient SIG54 in intra prediction, that is, the interpolation filter coefficient SIG8 calculated by the intra prediction value generation unit 20. For this reason, errors in intra prediction can be reduced, and as a result, compression performance can be further improved.

Also, the moving picture coding apparatus 1 applies a prediction error or a coding error as an error based on the input image SIG1 and the pre-intra prediction value SIG18 by the error minimizing filter coefficient calculation unit 22. For this reason, when applying prediction errors, it is not necessary to perform orthogonal transformation processing, quantization processing, inverse quantization processing, and inverse transformation processing, compared to the case of applying coding errors, so error minimization is achieved. The processing amount in the filter coefficient calculation unit 22 can be reduced. In addition, when coding errors are applied, compression performance is particularly improved in intra coding that generates a large amount of prediction residual code, high bit rate coding, etc., compared to the case where prediction errors are applied. Can be made.

Also, the moving picture encoding apparatus 1 can calculate the filter coefficient based on the Wiener filter method by the error minimizing filter coefficient calculation unit 22.

Also, in the video encoding device 1, the intra prediction value generation unit 20 classifies the intra prediction mode into m groups, and calculates an interpolation filter coefficient for each of the classified groups. For this reason, compared with the case where one filter coefficient common to all intra prediction modes is calculated, it is possible to improve the encoding performance by suppressing an increase in the amount of information regarding the filter coefficient while reducing the prediction error. it can.

Second Embodiment
FIG. 5 is a block diagram of a moving image processing system BB according to the second embodiment of the present invention. The moving image processing system BB is different from the moving image processing system AA according to the first embodiment of the present invention shown in FIG. 1 in that a moving image encoding device 1A is provided instead of the moving image encoding device 1, and a moving image The difference is that a moving image decoding apparatus 100A is provided instead of the decoding apparatus 100.

[Configuration and Operation of Moving Image Encoding Device 1A]
FIG. 6 is a block diagram of the video encoding apparatus 1A. The moving image encoding device 1A is different from the moving image encoding device 1 according to the first embodiment of the present invention shown in FIG. 2 in that an inter prediction value generation unit 10A is provided instead of the inter prediction value generation unit 10. The difference is that an intra prediction value generation unit 20A is provided instead of the intra prediction value generation unit 20. Note that in the moving image encoding device 1A, the same components as those in the moving image encoding device 1 are denoted by the same reference numerals, and description thereof is omitted.

Inter prediction value generation unit 10A includes a decimal accuracy MC prediction value generation unit 11 and an integer accuracy MC prediction value generation unit 12A.

The integer precision MC predicted value generation unit 12A receives the input image SIG1 and the locally decoded image SIG17 supplied from the local memory 60. The integer precision MC predicted value generation unit 12A generates a predicted value based on the integer precision MC using the input image SIG1 and the local decoded image SIG17, and then performs error minimization filter processing described later to obtain the integer precision MC. It outputs as the inter prediction value SIG7 based on it, and outputs the below-mentioned interpolation filter coefficient SIG21 and the control information SIG6 regarding the inter prediction of integer precision. The integer precision MC predicted value generation unit 12A will be described in detail below with reference to FIG.

FIG. 7 is a block diagram of the integer precision MC predicted value generation unit 12A. The integer accuracy MC predicted value generation unit 12 </ b> A includes a prior MC predicted value calculation unit 13, an error minimization filter coefficient calculation unit 14, and a filter processing unit 15.

The prior MC prediction value calculation unit 13 receives the input image SIG1 and the locally decoded image SIG17. The prior MC prediction value calculation unit 13 generates a prediction value by inter prediction based on integer accuracy MC using the input image SIG1 and the local decoded image SIG17, and outputs the prediction value as the prior inter prediction value SIG22. The control information SIG6 related to the inter prediction is output.

The error minimizing filter coefficient calculation unit 14 receives the input image SIG1 and the prior inter prediction value SIG22. The error minimizing filter coefficient calculation unit 14 calculates an interpolation filter coefficient that minimizes an error based on the input image SIG1 and the pre-inter prediction value SIG22, and outputs the interpolation filter coefficient SIG21.

The prediction error is a difference between the input image SIG1 and the prior inter prediction value SIG22 (an error generated by inter prediction). When calculating an interpolation filter coefficient that minimizes the prediction error, the error minimization filter coefficient calculation unit 14 calculates an interpolation filter coefficient that minimizes the difference between the input image SIG1 and the pre-inter prediction value SIG22. As a method for calculating the interpolation filter coefficient, for example, a calculation method based on the Wiener filter method can be used.

On the other hand, the coding error is an error obtained by combining the above-described prediction error and the quantization error generated in the quantization process. When calculating an interpolation filter coefficient that minimizes the coding error, the error minimizing filter coefficient calculation unit 14 first obtains a difference between the input image SIG1 and the prior inter prediction value SIG22. Next, after the same processing as the DCT / quantization unit 30 is performed on the obtained difference, the same processing as that of the inverse DCT / inverse quantization unit 50 is performed to obtain the above-described encoding error. Next, an interpolation filter coefficient that minimizes the obtained coding error is calculated. As a method for calculating the interpolation filter coefficient, for example, a calculation method based on the Wiener filter method can be used.

The filter processing unit 15 receives the interpolation filter coefficient SIG21 and the pre-inter prediction value SIG22 as inputs. The filter processing unit 15 performs an error minimizing filter process on the inter prediction value SIG22 using the interpolation filter coefficient SIG21, and outputs the predicted value after the filter process as an inter prediction value SIG7 based on integer precision MC. To do.

6, the intra predicted value generation unit 20A receives the input image SIG1 and the local decoded image SIG17 supplied from the local memory 60 as inputs. This intra prediction value generation unit 20A generates a prediction value based on intra prediction using the input image SIG1 and the local decoded image SIG17, outputs the prediction value as the intra prediction value SIG9, and outputs control information SIG10 related to intra prediction.

[Configuration and Operation of Video Decoding Device 100A]
FIG. 8 is a block diagram of the video decoding device 100A. The video decoding device 100A is different from the video decoding device 100 according to the first embodiment of the present invention shown in FIG. 4 in that an entropy decoding unit 110A is provided instead of the entropy decoding unit 110, and an inter prediction value generation unit. The difference is that the inter prediction value generation unit 130 </ b> A is provided instead of 130, and the point that the intra prediction value generation unit 140 </ b> A is provided instead of the intra prediction value generation unit 140. In the moving picture decoding apparatus 100A, the same components as those in the moving picture decoding apparatus 100 are denoted by the same reference numerals, and the description thereof is omitted.

The entropy decoding unit 110A receives the encoded data SIG2. The entropy decoding unit 110A analyzes the content described in the encoded data SIG2 according to the encoded data structure and performs entropy decoding, and performs the residual signal SIG51 and the interpolation filter coefficient SIG52 (interpolation filter coefficient SIG3 in the decimal precision MC). ), Control information SIG53, and interpolation filter coefficient SIG71 (interpolation filter coefficient SIG21) with integer precision MC are acquired and output.

The inter prediction value generation unit 130A includes a decimal accuracy MC prediction value generation unit 131 and an integer accuracy MC prediction value generation unit 132A.

The integer precision MC predicted value generation unit 132A includes control information SIG57 (control information SIG6 related to integer precision inter prediction) included in the control information SIG53, interpolation filter coefficient SIG71 in integer precision MC, The decoded image SIG64 supplied from the local memory 150 is used as an input. This integer precision MC predicted value generation unit 132A first refers to the decoded image SIG64 based on the control information SIG57 related to integer precision inter prediction, and generates a predicted value based on the integer precision MC in the processing block. Next, the prediction value based on the generated integer accuracy MC is subjected to filter processing using the interpolation filter coefficient SIG71 in the integer accuracy MC, and the prediction value after the filter processing is set as an inter prediction value SIG60 based on the integer accuracy MC. Output.

The intra predicted value generation unit 140A receives as input the control information SIG58 related to intra prediction (control information SIG10 related to intra prediction) included in the control information SIG53 and the decoded image SIG64 supplied from the local memory 150. The intra predicted value generation unit 140A refers to the decoded image SIG64 based on the control information SIG58 related to intra prediction, generates an intra predicted value in the processing block, and outputs the intra predicted value SIG61.

According to the above moving picture coding apparatus 1A and moving picture decoding apparatus 100A, the following effects can be obtained.

The moving image encoding apparatus 1A performs filtering processing on a frame before encoding, that is, a frame before encoding processing is performed by the filter processing unit 15. For this reason, the prediction performance that strongly affects the compression performance in predictive coding can be improved, and as a result, the compression performance can be improved.

Also, the moving image encoding apparatus 1A calculates an interpolation filter coefficient that minimizes an error based on the input image SIG1 and the pre-inter prediction value SIG22 by the error minimizing filter coefficient calculation unit 14, and the filter processing unit 15 calculates the error Filter processing is performed using the interpolation filter coefficient calculated by the minimizing filter coefficient calculation unit 14. Also, the moving picture decoding apparatus 100A uses the integer accuracy MC prediction value generation unit 132A to filter the interpolation filter coefficient SIG71 in integer accuracy MC, that is, the interpolation filter coefficient SIG21 calculated by the integer accuracy MC prediction value generation unit 12A. Process. For this reason, errors in motion compensation at integer pixel positions can be reduced, and as a result, compression performance can be further improved.

Also, the moving picture encoding apparatus 1A applies a prediction error or a coding error as an error based on the input image SIG1 and the pre-inter prediction value SIG22 by the error minimizing filter coefficient calculation unit 14. For this reason, when applying prediction errors, it is not necessary to perform orthogonal transformation processing, quantization processing, inverse quantization processing, and inverse transformation processing, compared to the case of applying coding errors, so error minimization is achieved. The processing amount in the filter coefficient calculation unit 14 can be reduced. In addition, when coding errors are applied, compression performance is particularly improved in intra coding that generates a large amount of prediction residual code, high bit rate coding, etc., compared to the case where prediction errors are applied. Can be made.

Also, the moving image encoding apparatus 1A can calculate the filter coefficient based on the Wiener filter method by the error minimizing filter coefficient calculation unit 14.

It should be noted that the processing of the moving image encoding device 1, 1A or the moving

image decoding device

100, 100A of the present invention is recorded on a computer-readable non-transitory recording medium, and the program recorded on this recording medium is recorded as a moving image. The present invention can be realized by causing the encoding apparatus 1 or 1A or the moving

image decoding apparatus

100 or 100A to read and execute the apparatus.

Here, for example, a nonvolatile memory such as an EPROM or a flash memory, a magnetic disk such as a hard disk, a CD-ROM, or the like can be applied to the above-described recording medium. Further, reading and execution of the program recorded on the recording medium is performed by a processor provided in the moving image encoding device 1, 1A or the moving

image decoding device

100, 100A.

In addition, the above-described program may be transmitted from the moving image encoding device 1 or 1A or the moving

image decoding device

100 or 100A storing the program in a storage device or the like via a transmission medium or by a transmission wave in the 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. Furthermore, what can implement | achieve the above-mentioned function in combination with the program already recorded on the moving image encoder 1, 1A or the moving

image decoder

100, 100A, what is called a difference file (difference program) may be sufficient.

As described above, the embodiments of the present invention have been described in detail with reference to the drawings. However, the specific configuration is not limited to the embodiments, and includes a design that does not depart from the gist of the present invention.

For example, the moving picture encoding apparatus 1 in the first embodiment described above is provided with the integer accuracy MC predicted value generation unit 12A in the second embodiment described above instead of the integer accuracy MC predicted value generation unit 12, and the first The moving picture decoding apparatus 100 according to the first embodiment may be provided with the integer accuracy MC predicted value generation unit 132A according to the second embodiment described above, instead of the integer accuracy MC predicted value generation unit 132.

Claims

A video encoding device that allows predictive encoding using an encoded local decoded image,
Pre-prediction value generation means for generating a prediction value by at least one of inter prediction based on integer precision motion compensation and intra prediction using the input image and the local decoded image;
An error minimizing filter coefficient calculating means for calculating a filter coefficient for minimizing an error based on the input image and the predicted value generated by the prior predicted value generating means;
Filter processing means for performing filter processing on the prediction value generated by the prior prediction value generation means using the filter coefficient calculated by the error minimization filter coefficient calculation means. Video encoding device.
The error minimizing filter coefficient calculation means includes
A prediction error that is a difference between the input image and the prediction value generated by the prior prediction value generation unit;
Encoding error obtained by performing orthogonal transformation processing, quantization processing, inverse quantization processing, and inverse transformation processing on the prediction error,
The moving picture encoding apparatus according to claim 1, wherein a filter coefficient that minimizes one of the two is calculated.
3. The moving picture coding apparatus according to claim 1, wherein the error minimizing filter coefficient calculating means calculates the filter coefficient based on a Wiener filter method.
The prior prediction value generation means generates a prediction value by intra prediction using the input image and the local decoded image,
4. The moving picture encoding apparatus according to claim 1, wherein the error minimizing filter coefficient calculating unit calculates the filter coefficient in accordance with an intra prediction mode. 5.
The error minimizing filter coefficient calculating means classifies the intra prediction modes into m groups (where m is a natural number smaller than the number of types of intra prediction modes and does not include zero), and for each classified group, The moving image encoding apparatus according to claim 4, wherein a filter coefficient is calculated.
A video decoding device that allows predictive decoding using a decoded image,
Entropy decoding means for entropy decoding encoded data to obtain filter coefficients;
A prediction value is generated using the decoded image by at least one of inter prediction based on integer-precision motion compensation and intra prediction, and the generated prediction value is acquired by the entropy decoding unit. And a filter processing means for performing filter processing using the filter coefficient.
The entropy decoding means performs entropy decoding on the encoded data to obtain a filter coefficient calculated according to the intra prediction mode,
The moving image decoding apparatus according to claim 6, wherein the filter processing unit generates a prediction value by intra prediction using the decoded image.
The intra prediction modes are classified in advance into m groups (where m is a natural number that is smaller than the number of types of intra prediction modes and does not include zero), and the filter coefficient is calculated for each classified group. Has been
The moving picture decoding apparatus according to claim 7, wherein the entropy decoding unit performs entropy decoding on the encoded data to obtain a filter coefficient calculated for each group.
A moving image processing system comprising: a moving image encoding device that encodes a moving image to generate encoded data; and a moving image decoding device that decodes encoded data generated by the moving image encoding device. ,
The moving image encoding device is:
Pre-prediction value generating means for generating a prediction value by at least one of inter prediction based on integer precision motion compensation and intra prediction using an input image and a local decoded image;
An error minimizing filter coefficient calculating means for calculating a filter coefficient for minimizing an error based on the input image and the predicted value generated by the prior predicted value generating means;
Encoding side filter processing means for performing filter processing on the prediction value generated by the prior prediction value generation means using the filter coefficient calculated by the error minimization filter coefficient calculation means;
Entropy encoding means for entropy encoding the filter coefficient calculated by the error minimizing filter coefficient calculation means to generate encoded data, and
The moving picture decoding device comprises:
Entropy decoding means for entropy decoding the encoded data generated by the entropy encoding means and obtaining the filter coefficient calculated by the error minimizing filter coefficient calculation means;
A prediction value is generated using a decoded image generated by inter prediction based on motion compensation of integer precision and intra prediction, which is used when the prediction value is generated by the prior prediction value generation unit, and is generated. And a decoding-side filter processing unit that performs a filter process on the predicted value using the filter coefficient acquired by the entropy decoding unit.
A moving picture coding method in a moving picture coding apparatus that includes a predictive value generating means, an error minimizing filter coefficient calculating means, and a filter processing means, and that allows predictive coding using a coded local decoded image There,
A first step of generating a prediction value by at least one of inter prediction based on integer-precision motion compensation and intra prediction, using the input image and the locally decoded image; When,
A second step in which the error minimizing filter coefficient calculating means calculates a filter coefficient that minimizes an error based on the input image and the predicted value generated in the first step;
The filter processing means includes a third step of performing filter processing on the prediction value generated in the first step using the filter coefficient calculated in the second step. A video encoding method.
A moving picture decoding method in a moving picture decoding apparatus that includes entropy decoding means and filter processing means and allows predictive decoding using decoded pictures,
A first step in which the entropy decoding means obtains a filter coefficient by entropy decoding the encoded data;
The filter processing means generates a prediction value using the decoded image by at least one of inter prediction based on integer-precision motion compensation and intra prediction, and for the generated prediction value, the first And a second step of performing a filter process using the filter coefficient acquired in the first step.
A moving picture coding method in a moving picture coding apparatus that includes a pre-prediction value generation means, an error minimization filter coefficient calculation means, and a filter processing means, and that allows predictive coding using a coded local decoded image A program for causing a computer to execute,
A first step of generating a prediction value by at least one of inter prediction based on integer-precision motion compensation and intra prediction, using the input image and the locally decoded image; When,
A second step in which the error minimizing filter coefficient calculating means calculates a filter coefficient that minimizes an error based on the input image and the predicted value generated in the first step;
The filter processing means causes the computer to execute a third step of performing filter processing on the prediction value generated in the first step using the filter coefficient calculated in the second step. Program for.
A program for causing a computer to execute a moving picture decoding method in a moving picture decoding apparatus that includes entropy decoding means and filter processing means and allows predictive decoding using decoded pictures,
A first step in which the entropy decoding means obtains a filter coefficient by entropy decoding the encoded data;
The filter processing means generates a prediction value using the decoded image by at least one of inter prediction based on integer-precision motion compensation and intra prediction, and for the generated prediction value, the first A program for causing a computer to execute a second step of performing a filtering process using the filter coefficient acquired in the first step.