JP6839246B2 - Video processing equipment, processing methods and computer-readable storage media - Google Patents

Description

The present invention relates to a moving image processing device, a processing method, and a computer-readable storage medium.

Patent Document 1 discloses intra-prediction (in-frame prediction) / inter-frame prediction (inter-frame prediction), residual transformation, entropy coding, and a moving image coding device and decoding device using an in-loop filter. ing. FIG. 6 is a configuration diagram of the coding device described in Patent Document 1. Note that the coding is performed in units of blocks that are one of a plurality of different sizes obtained by dividing the frame.

First, the input image data is input to the inter prediction unit 15 and the intra prediction unit 16. The inter-prediction unit 15 is input with the image data of the previous frame from the frame buffer 17, and the intra-prediction unit 16 is input with the image data of the already processed block of the same frame as the processing target from the addition unit 14. To. The inter-prediction unit 15 outputs a prediction block of a block to be processed by inter-frame prediction based on the previous frame. The intra prediction unit 16 outputs a prediction block of the processing target block based on another block in the same frame as the processing target block. Then, either the output of the inter-prediction unit 15 or the intra-prediction unit 16 is output to the subtraction unit 10 depending on whether the inter-frame prediction is applied to the processing target block or the intra-frame prediction is applied.

The subtraction unit 10 outputs an error (residual) signal indicating an error between the image of the block to be processed and the prediction image output by the inter prediction unit 15 or the intra prediction unit 16. The conversion / quantization unit 11 performs orthogonal conversion and quantization on the error signal and outputs a level value. The coding unit 12 entropy-codes this level value and side information (not shown) to generate a bit stream. The side information is information necessary for reconstructing the pixel value used in the decoding device, and is information such as a coding mode, a quantization parameter, and a block size indicating whether intra-prediction or inter-prediction is used. Includes.

The inverse quantization / inverse conversion unit 13 performs the reverse processing of the conversion / quantization unit 11 to generate an error signal. The addition unit 14 adds the error signal output by the inverse quantization / inverse conversion unit 13 and the prediction image output by the inter prediction unit 15 or the intra prediction unit 16 to generate a processing target block, and the intra prediction unit 16 Is output to the in-loop filter 18. When the in-loop filter 18 receives all the blocks of one frame, the in-loop filter 18 generates a locally decoded image corresponding to the frame and outputs the locally decoded image to the frame buffer 17. This locally decoded image is used for inter-frame prediction in the inter-prediction unit 15.

Non-Patent Document 1 discloses that the in-loop filter 18 performs deblocking processing and sample adaptation offset processing. The deblocking process is a process for reducing the distortion generated at the boundary of blocks in the frame. This prevents the image quality deterioration from propagating in the inter-frame prediction. The sample adaptive offset process is a process of adding / subtracting an offset value with respect to a pixel value.

FIG. 7 is a configuration diagram of the decoding device described in Patent Document 1. The bit stream generated by the encoding device is entropy-decoded by the decoding unit 20, and the level value and side information are taken out. The inverse quantization / inverse conversion unit 21 generates an error signal from the level value. This error signal is output by the inter-prediction unit 23 or the intra-prediction unit 24 depending on whether the block corresponding to the error signal is inter-frame predicted or intra-frame predicted. Is added by the prediction image of the above and the addition unit 22. As a result, the addition unit 22 reproduces the block. The block reproduced by the addition unit 22 is output to the intra prediction unit 24 for intra-frame prediction. Further, the block reproduced by the addition unit 22 is also output to the in-loop filter 26. When the in-loop filter 26 receives all the blocks of one frame, the in-loop filter 26 generates a locally decoded image corresponding to the frame and outputs the locally decoded image to the frame buffer 25. This locally decoded image is used for inter-frame prediction in the inter-prediction unit 23, and at the same time, is output as output image data.

FIG. 8 shows the configuration of the deblocking filter provided in the in-loop filters 18 and 26 disclosed in Non-Patent Document 1. Deblocking is performed for each of the vertical boundary and the horizontal boundary. In FIG. 8, the component to which the odd-numbered reference code is assigned deblocks the boundary in the vertical direction, and the component to which the reference code of the even number is given deblocks the boundary in the horizontal direction. First, the conversion block boundary detection unit 31 detects the vertical boundary of the conversion block based on the side information indicating the size of the conversion block. Subsequently, the prediction block boundary detection unit 33 detects the vertical boundary of the prediction block based on the side information indicating the size of the prediction block. The conversion block is a block related to orthogonal transformation executed by the conversion / quantization unit 11, and the prediction block is a block in the prediction processing performed by the inter-prediction unit 15 or the intra-prediction unit 16. In the following, the boundaries of the conversion block and the prediction block are collectively referred to as "boundaries".

The boundary strength determination unit 35 determines the side information, specifically, whether it is an intra prediction or an inter prediction, a boundary of a conversion block, and whether or not a non-zero orthogonal conversion coefficient exists. The boundary strength is set to 0 depending on whether the difference between the motion vectors of the two blocks sandwiched is equal to or greater than the threshold value, the reference image of the motion compensation of the two blocks sandwiching the boundary is different, or the number of motion vectors is different. Evaluate on a three-point scale of 1, 2. It is assumed that the boundary strength 0 is the weakest and the boundary strength 2 is the strongest. The filter determination unit 37 determines the boundary of the processing target based on the determination criteria using the boundary strength of the boundary of the processing target, the quantization parameter included in the side information, and the pixel value of the pre-deblocking image which is the filter target image. Whether or not to apply a filter to, and if a filter is applied, whether to apply a weak filter or a strong filter is determined. The filter unit 39 applies a filter to the pre-deblocking image according to the determination of the filter determination unit 37 to perform deblocking.

The processing by the conversion block boundary detection unit 32, the prediction block boundary detection unit 34, the boundary strength determination unit 36, and the filter determination unit 38 is performed by the conversion block boundary detection unit 31, the prediction block boundary detection unit 33, the boundary strength determination unit 35, and the filter. Only the direction of the target boundary is different from the processing in the determination unit 37, and the description thereof will be omitted again. Then, the filter unit 40 applies a filter to the filtered target image, which is an image obtained by applying the filter to the vertical boundary output by the filter unit 39, in accordance with the determination of the filter determination unit 38, and outputs the deblocked image. To do.

FIG. 9 shows the vertical boundaries. In addition, each of pxy and qxy (x and y are integers of 0 to 3 respectively) is a pixel. In Non-Patent Document 1, a total of 16 pixel values of px0, qx0, px3, and qx3 (where x is an integer from 0 to 3) are used to determine the filter type. Further, the filtering process is performed in units of a total of 32 pixels of pxy and qxy (x and y are integers of 0 to 3 respectively). That is, in the case of the boundary in the vertical direction, whether or not to apply the filter in units of 4 pixels in the vertical direction and, in the case of applying the filter, the intensity thereof are determined. In the case of a weak filter, the pixel values of p0y and q0y are changed by the filter processing, and in the case of a strong filter, the pixel values of p2y, p1y, p0y, q0y, q1y and q2y are changed by the filter processing. Will be done. In the case of the boundary in the horizontal direction, FIG. 9 is rotated by 90 degrees to perform the same processing as the boundary in the vertical direction. That is, in the case of the boundary in the horizontal direction, the relationship between the position of the pixel used for determining the filter type and the position of the pixel whose value is changed by the filter and the boundary is the same as in the case of the boundary in the vertical direction.

Patent Document 2 discloses a configuration in which the boundary strength becomes stronger as the predicted block becomes larger in order to suppress the block distortion. Further, Non-Patent Document 2 discloses that in order to suppress block distortion, coding control is performed so that a block having a large size is difficult to be selected.

Japanese Unexamined Patent Publication No. 2014-197847 Japanese Unexamined Patent Publication No. 2011-223302

ITU-T H.265 High Efficiency Video Coding JCTVC-L0232 AHG6: On deblocking filter and parameters signaling

Patent Document 2 considers only the prediction block, and block distortion occurs depending on the size of the conversion block. Further, in the configuration of Non-Patent Document 2, the amount of code increases.

The present invention provides a moving image processing device, a processing method, and a computer-readable storage medium that reduce block distortion without increasing the amount of code.

According to one aspect of the present invention, the moving image processing device is a moving image processing device, and determines whether or not to apply a filter to a detecting means for detecting a block boundary and a boundary detected by the detecting means. A determination means for determining is provided, wherein the boundary is the boundary of the conversion block, and the horizontal or vertical size of at least one conversion block among the conversion blocks on both sides of the boundary. there If it is smaller than the first threshold value, determines to apply a strong filter to the boundary, a processing apparatus, the determining means, when applying the filter to the boundary, on the border of the three or more different intensity filter The filter to be applied is determined, and the boundary is the boundary of the conversion block, and the horizontal or vertical size of at least one conversion block among the conversion blocks on both sides of the boundary is equal to or larger than the first threshold value. , It is determined that the strongest filter among the three or more types of filters having different intensities is applied to the boundary, while when the size is less than the first threshold value, the three or more types are applied to the boundary. It is characterized in that any of the remaining two or more types of filters having different intensities other than the strongest filter among the filters having different intensities is applied.

According to one aspect of the present invention, the moving image processing method in the processing device is a moving image processing method in the processing device, in which a detection step for detecting a block boundary and a filter for the boundary detected in the detection step are applied. Including a determination step of determining whether or not to apply, in the determination step, the boundary is the boundary of the conversion block, and the horizontal direction of at least one conversion block among the conversion blocks on both sides of the boundary. or the vertical size of is the first threshold value or more, determines to apply a strong filter to the boundary, a processing method, in the determination step, when applying a filter to the boundary, three or more different intensities The filter to be applied to the boundary is determined from the filter of the above, the boundary is the boundary of the conversion block, and the horizontal or vertical size of at least one conversion block among the conversion blocks on both sides of the boundary is the first. If it is 1 threshold or more, it is determined that the strongest filter among the three or more different intensity filters is applied to the boundary, while if the size is less than the first threshold, the boundary is applied. In addition, any of the remaining two or more types of filters having different intensities other than the strongest filter among the three or more types of filters having different intensities is applied .

According to the present invention, the block distortion can be reduced without increasing the code amount.

The block diagram of the deblocking filter according to one Embodiment. The block diagram of the deblocking filter according to one Embodiment. The figure which shows the filter coefficient of a strong filter. The figure which shows the filter coefficient by one Embodiment. Explanatory drawing of the filter processing by one Embodiment. The block diagram of the coding apparatus according to one Embodiment. The block diagram of the decoding apparatus according to one Embodiment. The block diagram of the deblocking filter. The explanatory view of the pixel used for the filter determination in the case of the boundary in the vertical direction, and the pixel to be filtered.

Hereinafter, exemplary embodiments of the present invention will be described with reference to the drawings. The following embodiments are examples, and the present invention is not limited to the contents of the embodiments. Further, in each of the following figures, components that are not necessary for the description of the embodiment will be omitted from the drawings.

<First Embodiment>
The basic configurations of the coding device and the decoding device in this embodiment are the same as those in FIGS. 6 and 7, and the description thereof will be omitted again. Further, the configurations of the in-loop filters 18 and 26 of the coding device and the decoding device according to the present embodiment are the same. Hereinafter, the deblocking filter included in the in-loop filters 18 and 26 of the coding device and the decoding device according to the present embodiment will be described, but the moving image processing will be collectively performed without distinguishing between the coding device and the decoding device. It shall be called a device. FIG. 1 shows a deblocking filter included in the in-loop filters 18 and 26. In the deblocking filter shown in FIG. 1, the same reference numerals are given to the same components as the deblocking filter shown in FIG. 8, and the description thereof will be omitted again. Further, the deblocking filter according to the present embodiment also has a portion for filtering the vertical boundary (odd reference number) and a portion for filtering the horizontal boundary, similarly to the deblocking filter shown in FIG. Although there are (even reference codes), the processing contents are the same except for the direction of the boundary, so only the vertical boundary will be described below.

In the present embodiment, the boundary strength determination unit 55 uses the conversion block size for determining the boundary strength in addition to the conventional boundary strength determination criteria shown in FIG. First, the conventional criteria for determining the boundary strength will be described. The boundary strength determination unit 55 determines that the boundary strength is 2 when one block of the vertical boundary shown in FIG. 9 is an intra prediction. Further, the boundary strength determination unit 55 determines that the boundary strength is 1 when the boundary is the boundary of the conversion block and the non-zero orthogonal conversion coefficient exists. Further, the boundary strength determining unit 55 determines that the boundary strength is 1 when the absolute value of the difference between the motion vectors of the two blocks sandwiching the boundary is one pixel or more. Further, the boundary strength determining unit 55 determines that the boundary strength is 1 if the reference images of the motion compensation of the two blocks sandwiching the boundary are different or the number of motion vectors is different. Then, the boundary strength determination unit 55 determines that the boundary strength is 0 if none of the above applies.

In the present embodiment, the boundary strength determination unit 55 first determines that the vertical boundary shown in FIG. 9 is the boundary of the conversion block and the vertical size of at least one conversion block is equal to or larger than the first threshold value. 1 is added to the boundary strength determined by the conventional determination criteria described in the above, and this is output as the final boundary strength. That is, in the present embodiment, the boundary strength determination unit 55 outputs any value of the boundary strengths 0 to 3. For example, the first threshold value can be 16 pixels in which the block distortion of the conversion block becomes conspicuous.

Whether the filter determination unit 37 applies the filter to the boundary according to the determination criteria of the filter based on the boundary strength of the boundary, the quantization parameter included in the side information, and the pixel value of the image before deblocking. Whether or not to apply the filter, and when applying the filter, decide whether to apply the weak filter or the strong filter.

Here, in the determination criterion, the stronger the boundary strength value, the higher the probability of determining that the filter is applied. More specifically, regarding the luminance value, it is determined that the filter is not applied when the boundary intensity is 0. On the other hand, when the boundary strength is 1 or more, the value calculated from the pixel values of a total of 12 pixels of px0, qx0, px3, qx3 (where x is an integer from 0 to 2) constitutes the boundary 2 Whether or not to apply the filter is determined by comparing the values obtained from the mean values of the quantization parameters of one block. Further, regarding the color difference value, it is determined that the filter is applied when the boundary strength is 2 or more. In other words, in the determination criterion, one of the conditions for applying the filter is that the value of the intensity of the boundary is equal to or higher than the threshold value (1 for the luminance value and 2 for the color difference value). When the boundary intensity value is equal to or greater than the threshold value, the brightness value is further determined from the pixel values of a total of 12 pixels of px0, qx0, px3, qx3 (where x is an integer from 0 to 2). If the calculated value is less than the value obtained from the average value of the quantization parameters of the two blocks constituting the boundary, it is determined that the filter is applied. Regarding the color difference value, if the boundary intensity value is equal to or greater than the threshold value, it is determined that the filter is applied.

In addition, in the determination criteria, the stronger the boundary strength, the higher the probability of determining that a strong filter is applied. Specifically, when applying a filter to the luminance value, 6 values are calculated based on the pixel values of a total of 16 pixels of px0, qx0, px3, qx3 (where x is an integer from 0 to 3). However, if all the values meet the predetermined criteria, it is decided to apply the strong filter. Two of the predetermined criteria are satisfied that the difference between p00 and q00 and the difference between p03 and q03 are less than the second threshold, and the second threshold is set based on the boundary strength. More specifically, the intermediate value is obtained based on the boundary strength. This intermediate value is set to a larger value as the boundary strength becomes stronger. Then, the second threshold value is obtained based on this intermediate value. The relationship between the intermediate value and the second threshold value is predetermined. The relationship between the intermediate value and the second threshold value is such that when the intermediate value is the first value, the second threshold value is the third value, and when the intermediate value is the second value larger than the first value, the second value is obtained. When the 2 threshold value is the 4th value, the 4th value is set to be a value equal to or higher than the 3rd value. That is, when the boundary strength changes from the first value to a larger second value, the second threshold value changes from the third value to a value equal to or higher than the third value. Then, in the determination criterion, one of the conditions for selecting a strong filter is that the absolute value of the difference between p00 and q00 and the absolute value of the difference between p03 and q03 are smaller than the second threshold value. Therefore, the stronger the boundary strength, the higher the probability that the strong filter will be applied. One type of filter is used for the color difference value.

As described above, in the present embodiment, the size of the conversion block is used for determining the boundary strength. More specifically, the boundary of the prediction block and the boundary of the conversion block are detected, and the size of the conversion block is used to determine the strength of the boundary of the conversion block. At this time, if the size of at least one of the conversion blocks on both sides of the boundary in the same direction as the boundary is equal to or larger than the first threshold value, it is determined that the strength of the boundary is stronger than when it is less than the first threshold value. To do. For example, when the size of at least one of the conversion blocks on both sides of the boundary in the same direction as the boundary is equal to or larger than the first threshold value, the strength of the boundary is set to a predetermined value, for example, from the case where the size is less than the first threshold value. Increase by 1. The filter determination units 37 and 38 determine whether or not to apply the filter to the boundary based on the determination criteria including the intensity of the boundary. As described above, the stronger the intensity of the boundary, the more the filter is applied to the boundary. The probability of being applied is high. Therefore, when the size of the conversion block is equal to or larger than the first threshold value, the probability that the filter is applied is high and the filter distortion can be suppressed from being conspicuous. Further, since the block size is not suppressed to increase, the code amount does not increase.

Further, when applying a filter to a boundary, the filter determination units 37 and 38 determine a filter to be applied to the boundary from a plurality of filters having different intensities. In the above-described embodiment, the plurality of filters having different intensities are two types, a weak filter and a strong filter having a stronger filter intensity than the weak filter. However, it is also possible to use three or more filters having different intensities. Then, the filter determination units 37 and 38 set a second threshold value for determining the filter to be applied to the boundary based on the strength of the boundary. Here, in the determination criteria, the stronger the boundary strength, the larger the second threshold value, and the larger the second threshold value, the higher the possibility that a stronger filter will be selected. Therefore, when the size of the conversion block is equal to or larger than the first threshold value, the probability that a stronger filter is applied increases, and it is possible to suppress the conspicuous filter distortion.

<Second embodiment>
Subsequently, the second embodiment will be described focusing on the differences from the first embodiment. In the present embodiment, the boundary strength determination unit 55 and the filter determination unit 37 of FIG. 1, and the boundary strength determination unit 56 and the filter determination unit 38 are connected to the boundary strength determination unit 60 and the filter determination unit 61 shown in FIG. 2, respectively. Replace. The processing by the boundary strength determination unit 60 is the same as that of the boundary strength determination unit 35 in FIG. That is, the boundary strength determining unit 60 outputs the boundary strengths 0 to 2. In the present embodiment, when the boundary is the boundary of the conversion block, the filter determination unit 61 determines whether or not the block size in the corresponding direction (vertical or horizontal) is equal to or larger than the first threshold value. Then, if it is equal to or higher than the first threshold value, it is determined that the strong filter is applied. On the other hand, if the size is less than the first threshold value, it is determined by the conventional method whether or not the filter is applied, and if it is applied, whether to apply the strong filter or the weak filter. The first threshold value can be 16 pixels as in the first embodiment.

As described above, in the present embodiment, the strong filter is always applied when the size of the conversion block on any side of the boundary is equal to or larger than the first threshold value. Therefore, block distortion can be suppressed even if the conversion block size is large. In this embodiment as well, the strength of the filter can be set to 3 or more. In this case, when the conversion block size is equal to or larger than the first threshold value, the strongest filter is always applied.

<Third Embodiment>
Subsequently, the third embodiment will be described focusing on the differences from the second embodiment. The configuration of this embodiment is the same as that of the second embodiment. However, in the second embodiment, the strong filter is always applied when the boundary is the boundary of the conversion block and the size is equal to or larger than the first threshold value. As described above, the strong filter is applied to the pixels of pxy and qxy (where x is 0 to 2 and y is 0 to 3) in FIG. That is, the filter processing is performed on the pixels whose distance from the boundary is 3 pixels or less. FIG. 3 shows the filter coefficient when applying the strong filter. For example, the pixel p2y is changed based on the pixel values of the pixels p3y, p2y, p1y, p0y and q0y. In the present embodiment, when the boundary is the boundary of the conversion block and its size is equal to or larger than the threshold value, another type of filter is used in which the distance from the boundary of the pixel range to which the filter is applied is wider than that of the normal strong filter. To do. For example, if the boundary is the boundary of the conversion block and its size is equal to or larger than the threshold value, the applicable range is extended from the boundary to 7 pixels. FIG. 4 shows an example of the filter coefficient of another type of filter, and in FIG. 4, the filter target is from the boundary to 7 pixels.

As described above, in the present embodiment, when the conversion block size is equal to or larger than the first threshold value, the pixel range of the filter application based on the boundary is expanded as compared with the normal strong filter. With this configuration, block distortion can be suppressed even if the conversion block size is large.

<Fourth Embodiment>
In the first to third embodiments, the filter was similarly applied to both sides of the boundary. In the present embodiment, when the size of one conversion block of the boundary is equal to or larger than the first threshold value, for example, 16 pixels or more, the conversion block of the one boundary is subjected to any one of the first to third embodiments. The filter is determined according to the process of, and the filter determined by the conventional method is applied to the other block. For example, as shown in FIG. 5, the size of the conversion block on the right side of the boundary to be processed is equal to or larger than the first threshold value, and the size of the conversion block on the left side of the boundary is smaller than the first threshold value. When the method of the first embodiment is applied, the boundary strength determination unit 55 uses the first strength for determining the presence / absence of a filter for the pixels of the conversion block on the left side of the boundary and the filter type, and the pixels of the conversion block on the right side of the boundary. The presence or absence of a filter for the filter and the second intensity for determining the filter type are determined respectively. According to the first embodiment, the second intensity is a value obtained by adding 1 to the first intensity. Further, when the method of the second embodiment is applied, the filter determination unit 61 determines the presence / absence of a filter and the filter type for the pixels of the conversion block on the left side of the boundary according to the method of the prior art, and determines the pixels of the conversion block on the right side of the boundary. Decide to apply a strong filter to. Further, when the method of the third embodiment is applied, the filter determination unit 61 determines the presence / absence of a filter and the filter type for the pixels of the conversion block on the left side of the boundary according to the method of the prior art, and determines the pixels of the conversion block on the right side of the boundary. It is decided to apply a filter that extends the distance range from the boundary.

As a result, the filter strength to be applied and the range thereof can be limited to the conversion blocks of the first threshold value or more, and the increase in the filter processing load can be suppressed.

The processing device according to the present invention, that is, the coding device or the decoding device, can be realized by a program that operates the computer as the processing device. These computer programs are stored in a computer-readable storage medium or can be distributed over a network.

31, 32: Conversion block detection unit, 33, 34: Prediction block detection unit, 55, 56: Boundary strength determination unit, 37, 38: Filter determination unit

Claims (6)

It is a moving image processing device
Detection means to detect block boundaries and
It is provided with a deciding means for deciding whether or not to apply a filter to the boundary detected by the detecting means.
When the boundary is the boundary of the conversion block and the horizontal or vertical size of at least one conversion block among the conversion blocks on both sides of the boundary is equal to or larger than the first threshold value, the determination means . Is a processing device that decides to apply a strong filter to
The determination means is
When applying a filter to a boundary, determine the filter to be applied to the boundary from three or more types of filters with different intensities.
When the boundary is the boundary of the conversion block and the horizontal or vertical size of at least one conversion block among the conversion blocks on both sides of the boundary is equal to or larger than the first threshold value, the boundary is set to the above 3 Decided to apply the strongest of the different strength filters over different types,
On the other hand, when the size is less than the first threshold value, at the boundary, among the remaining two or more types of filters having different intensities other than the strongest filter among the three or more types of filters having different intensities. Apply one of the filters,
A processing device characterized by the fact that. The determining means is when the boundary is the boundary of the conversion block and the horizontal or vertical size of at least one conversion block among the conversion blocks on both sides of the boundary is equal to or larger than the first threshold value. The processing apparatus according to claim 1, wherein the filter applied to the boundary increases the distance from the boundary of the target pixel region as compared with the case where the threshold value is less than the first threshold value. In the determination means, the boundary is the boundary of the conversion block, and the horizontal or vertical size of one of the conversion blocks on both sides of the boundary is equal to or larger than the first threshold value, and the other. When the horizontal or vertical size of the conversion block is less than the first threshold value, the filter applied to the pixels in the one conversion block and the filter applied to the pixels in the other conversion block are different. , A processing device,
The different filters are filters in which the number of pixels referenced at the same time when applying the filter is different.
The processing apparatus according to claim 1 or 2. In the determination means, the boundary is the boundary of the conversion block, and the horizontal or vertical size of one of the conversion blocks on both sides of the boundary is equal to or larger than the first threshold value, and the other. When the horizontal or vertical size of the conversion block is less than the first threshold value, the filter applied to the pixels in the one conversion block and the filter applied to the pixels in the other conversion block are different. , A processing device,
The different filters are those in which the filter applied to the conversion block whose horizontal or vertical size is equal to or larger than the first threshold value is the strongest filter among the three or more types of filters having different intensities. The filter applied to the conversion block whose size is less than the first threshold value is among the remaining two or more types of filters having different intensities other than the strongest filter among the three or more types of filters having different intensities. Is one of the filters of
The processing apparatus according to claim 1 or 2. It is a method of processing moving images in a processing device.
A detection step that detects block boundaries and
Including a determination step of deciding whether or not to apply a filter to the boundary detected in the detection step.
In the determining step, the border is a border of the conversion block, and the horizontal or vertical size of the at least one translation block of the both sides of the conversion block of the boundary is smaller than the first threshold value, the A processing method that determines to apply a strong filter to the boundary,
In the determination step,
When applying a filter to a boundary, determine the filter to be applied to the boundary from three or more types of filters with different intensities.
When the boundary is the boundary of the conversion block and the horizontal or vertical size of at least one conversion block among the conversion blocks on both sides of the boundary is equal to or larger than the first threshold value, the boundary is set to the above 3 Decided to apply the strongest of the different strength filters over different types,
On the other hand, when the size is less than the first threshold value, at the boundary, among the remaining two or more types of filters having different intensities other than the strongest filter among the three or more types of filters having different intensities. Apply one of the filters,
A processing method characterized by that. A computer-readable storage medium comprising a program that causes a computer to function as the processing device according to any one of claims 1 to 4.

Images