Apparatus and method for deblocking filter processing

The present invention relates to a deblock filter processing apparatus and a deblock filter processing method applied to advanced multimedia data coding, particularly video coding using motion prediction based on a variable block size.

Generally, in video compression technology, various filters are used to improve image quality and compression rate. Images obtained by decoding low-bit rate compressed video often have block artifacts due to quantization noise and motion compensation. One of the main roles of the deblock filter processing device (hereinafter referred to as "deblock filter" or simply "filter") is to smooth the block boundaries in the decoded image and reduce or eliminate block artifacts. In addition, the deblock filter includes, for example, a post filter that removes noise when playing back an image on the decoder side and stabilizes the image quality to improve the image quality, and for example, noise when compressing the image on the encoder side. There is a loop filter that improves the image quality by removing it and improving the compression efficiency.

Conventionally, as a deblocking filter for improving image quality, there is one described in Patent Document 1. This document discloses a post-filter that applies filter strength to a decoded image according to the coding mode. In addition, a loop filter applied to both a reference image and a non-reference image has been proposed in order to improve the image quality of the decoded image.<patcit num="1"><text>Japanese Unexamined Patent Publication No. 2001-224031</text></patcit>

<p> However, the conventional deblock filter is applied to an image on a fixed block size basis.</p><p> For example, in ISO / IEC 14496 Part 10 of the video coding standard, DCT (discrete cosine transform) is applied to 4x4 size blocks (hereinafter, blocks with NxN size are abbreviated as "NxN blocks"). Traditional deblock filters are designed to be applied to boundaries of 4x4 blocks. Also, for example, in standard ISO / IEC14496 Part 2, the DCT is executed in 8x8 size, so conventional deblock filters are designed to be applied to the boundaries of 8x8 blocks. Such a filter design is good for taking DCT into account. This is because most of the block noise at low bit rates is generated by the DCT.</p><p> However, the above filter design is not optimal for the interframe wavelet video coding method, which has been attracting attention in recent years as a standard for next-generation video coding. Generally, in this method, MCTF (Motion Compensated Temporal Filtering) with block-based motion estimation / compensation is used in the time direction, and 2D-DWT (discrete wavelet transform) is used for spatial transform. Unlike the DCT, the DWT has the characteristic that it does not cause block artifacts in the decoded image. Therefore, in this method, the main cause of causing block artifacts is motion estimation / compensation. Inaccurate motion estimation and quantization can cause block artifacts, especially when performing motion estimation / compensation at low bit rates or in low latency modes with a small GOP (Group Of Picture) size. In the case of MCTF, this is done at each time decomposition level, so block artifacts can accumulate across all time decomposition levels.</p><p> Hereinafter, a general example of wavelet coding (MC wavelet coding) using motion estimation will be described with reference to FIG. Here, the case of GOP size = 8 will be described as an example.</p><p> As shown in Figure 9, there are eight original frames at level 0. After time-resolved processing with motion estimation and a wavelet filter in the time direction (Haar filter is shown in this example, but longer filters are also applicable), at level 1, low frequency frames L1, L2, A group consisting of L3 and L4 and a group consisting of high-frequency frames H1, H2, H3 and H4 are generated. Then, motion estimation and time filtering are further applied to the low-frequency frames L1, L2, L3, and L4, and at level 2, the group consisting of the low-frequency frames LL1 and LL2 and the high-frequency frames LH1 and LH2 are used. Is generated. Then, motion estimation and time filtering are performed again, and finally, at level 3, only two frames LLL1 and LLH1 exist.</p><p> Then, the frames LLL1, LLH1 at level 3, LH1, LH2 at level 2, and H1, H2, H3, and H4 at level 1 are subjected to spatial wavelet decomposition, followed by scanning and space. Entropy coding (variable length coding) is performed in consideration of scalability, time scalability, image quality scalability, etc., and a scalable stream is generated.</p><p> As is generally understood, the motion estimates used in MC wavelet coding are not always based on fixed block sizes such as ISO / IEC 13818 Part 2 16x16 and ISO / IEC 14496 Part 2 up to 16x16. This size can vary from as small as 4x4 to as large as 64x64, and even larger depending on the nature of the image. FIG. 10 illustrates variable block sizes (eg, 4x4 to 64x64) that can be used in block matching for motion estimation / compensation between reference frame A and current frame B.</p><p> Further, FIG. 11 is a diagram showing an application example of the conventional fixed block size-based deblock filter processing. Here, an example in which the deblocking filter processing shown by the broken line is applied to the frame having the block shown by the solid line in FIG. 11 will be described. Blocks S1, S2, and S3 each have a size selected for motion estimation, block S1 is 64x64 blocks, block S2 is 32x32 blocks, and block S3 is 16x16 blocks.</p><p> If the block size of motion estimation is 64x64 like block S1, there should be no block artifacts in this 64x64 block. In such a case, if a fixed size deblock filter process having a size smaller than 64x64 is applied, the power of the processing device (for example, CPU) is unnecessarily consumed. Not only that, important information is filtered, and as a result, the sharpness of the image is lost and the image becomes blurry, and in the end, there is a problem that high image quality cannot be realized.</p><p> The present invention has been made in view of this point, and provides a deblock filter processing device and a deblock filter processing method capable of achieving high image quality without consuming unnecessary power of the processing device. The purpose is.</p>

<p> The deblocking filter processing apparatus of the present invention is an acquisition means for acquiring a variable size motion estimation block in a frame subjected to motion estimation processing, and an application in which the deblocking filter processing is applied to the frame according to the acquired motion estimation block. Take a configuration with means and.</p><p> According to this configuration, the deblocking filter processing is applied to the frame according to the variable size motion estimation block in the frame subjected to the motion estimation processing. Therefore, for example, the block size in the deblocking filtering process is set to the size of the motion estimation block. When the size is adaptively changed, the block size of the deblock filter process and the size of the motion estimation block can be matched, and not only can the increase in the amount of the deblock filter process be suppressed, but also the increase in the amount of the deblock filter process can be suppressed. It is possible to prevent the sharpness from being lost unnecessarily, and it is possible to realize high image quality without consuming unnecessary power of the processing device.</p><p> In the above configuration, the deblocking filter processing apparatus of the present invention uses only the boundary between one motion estimation block in the frame and another motion estimation block adjacent to the motion estimation block, and the deblocking filter. Adopt a configuration to apply the process.</p><p> According to this configuration, the application position of the deblocking filter processing can be limited to only the boundary between the two motion estimation blocks, and the block size of the deblocking filtering process and the size of the motion estimation block can be matched. Not only can the increase in the amount of deblocking filtering be suppressed, but the sharpness of the image can be prevented from being wasted, and the image quality can be improved without consuming unnecessary power of the processing device. It can be realized.</p><p> In the above configuration, in the deblocking filter processing apparatus of the present invention, the application means determines the tap length of the deblocking filtering processing for the frame based on at least one of the coding information and the transmission information corresponding to the frame. Take the configuration to set.</p><p> According to this configuration, the tap length of the deblocking filter processing for the frame is set based on at least one of the coding information and the transmission information corresponding to the frame to which the motion estimation processing is performed, so that the deblocking filter is used. It is possible to efficiently set the tap length when applying the process.</p><p> In the above configuration, in the deblocking filter processing apparatus of the present invention, the application means sets the intensity of deblocking filtering for the frame based on at least one of the coding information and the transmission information corresponding to the frame. Take the configuration to do.</p><p> According to this configuration, the intensity of the deblocking filter processing for the frame is set based on at least one of the coding information and the transmission information corresponding to the frame to which the motion estimation processing is performed, so that the deblocking filtering processing is performed. It is possible to efficiently set the strength when applying.</p><p> In the above configuration, in the deblocking filter processing apparatus of the present invention, the application means applies the number of pixels to which the deblocking filter processing is applied to the frame based on at least one of the coding information and the transmission information corresponding to the frame. Take the configuration to set.</p><p> According to this configuration, the number of pixels to which the deblocking filter processing is applied to the frame is set based on at least one of the coding information and the transmission information corresponding to the frame to which the motion estimation processing is performed. The number of applied pixels when applying the filter processing can be efficiently performed.</p><p> The video coding apparatus of the present invention adopts a configuration having the above-mentioned deblock filter processing apparatus.</p><p> According to this configuration, the same operation and effect as the above-mentioned deblock filter processing device can be realized in the video coding device.</p><p> The video decoding apparatus of the present invention adopts a configuration having the above-mentioned deblocking filter processing apparatus.</p><p> According to this configuration, the same operation and effect as the above-mentioned deblock filter processing device can be realized in the video decoding device.</p><p> In the above configuration, the video decoding apparatus of the present invention adopts a configuration in which the time decomposition level to which the deblock filter processing should be applied in the frame is changed according to the signal transmitted from the corresponding video coding apparatus.</p><p> According to this configuration, it is possible to change the time decomposition level to which the deblocking filter processing should be applied and the number thereof when, for example, a predetermined instruction is transmitted from the corresponding video coding device, and the decoding in the video decoding device can be changed. It is possible to improve the efficiency of block filtering and reduce the load.</p><p> In the above configuration, the video decoding apparatus of the present invention adopts a configuration in which the deblock filter process is applied when the frame is reconstructed.</p><p> According to this configuration, the deblocking filter processing can be applied to the reconstructed frame.</p><p> The deblocking filter processing method of the present invention includes an acquisition step of acquiring a variable-sized motion estimation block in a frame to which motion estimation processing has been performed, and an application in which the deblocking filter processing is applied to the frame according to the acquired motion estimation block. To have a step and.</p><p> According to this method, the deblocking filter processing is applied to the frame according to the variable size motion estimation block in the frame subjected to the motion estimation processing. Therefore, for example, the block size in the deblocking filtering process is set to the size of the motion estimation block. When the size is adaptively changed, the block size of the deblock filter process and the size of the motion estimation block can be matched, and not only can the increase in the amount of deblock filter process be suppressed, but also the increase in the amount of the deblock filter process can be suppressed. It is possible to prevent the sharpness from being lost unnecessarily, and it is possible to realize high image quality without consuming unnecessary power of the processing device.</p>

<p> As described above, according to the present invention, it is possible to realize high image quality without unnecessarily consuming the electric power of the processing device.</p>

The gist of the present invention is to apply the deblock filter process to the frame according to the variable size motion estimation block in the frame to which the motion estimation process has been performed.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

(Embodiment 1) FIG. 1 is a block diagram showing a configuration of a video coding apparatus having a loop filter according to the first embodiment of the present invention.

The video coding apparatus 100 shown in FIG. 1 includes an image input unit 110, a motion estimation unit 120, a time filter 130, a spatial wavelet decomposition unit 140, a scanning / entropy encoding unit 150, a local decoding unit 160, a loop filter 170, and a reference. It has a frame buffer 180.

The image input unit 110 groups a predetermined number (fixed number or variable number) of adjacent frames in the input video sequence as 1 GOP, and then outputs the frames to the motion estimation unit 120. Further, the image input unit 110 may directly output the frame to the spatial wavelet decomposition unit 140 in order to obtain a coded frame independent of other frames for the purpose of random access or error recovery, for example.

The motion estimation unit 120 refers to the reference frame temporarily stored in the reference frame buffer 180, and performs motion estimation and motion compensation for the frame from the image input unit 110 within one GOP or between a plurality of GOPs. Do.

The time filter 130 performs time wavelet decomposition on the motion-compensated frame to generate low-frequency and high-frequency time frames at a plurality of time-resolved levels.

The spatial wavelet decomposition unit 140 performs spatial wavelet decomposition on the time frame from the time filter 130 or the frame from the image input unit 110.

The scanning / entropy coding unit 150 applies scanning and entropy coding to the frame from the spatial wavelet decomposition unit 140. The frame subjected to these processes is output as an encoded scalable stream.

The local decoding unit 160 locally decodes the frame from the spatial wavelet decomposition unit 140.

The loop filter 170, which is a feature of the present invention, performs the deblocking filter processing described later on the locally decoded frame except for the independent coded frame. The deblock filter process is performed corresponding to each of the plurality of time decomposition levels. When executing the deblocking filter processing, the loop filter 170 acquires the coding / transmission information in order to execute the deblocking filtering processing adaptively. The acquired coding / transmitting information includes motion estimation information related to motion estimation by the motion estimation unit 120, time decomposition information related to time wavelet decomposition by the time filter 130, and other quantization parameters, bit rate related information, and color components. Information such as required spatial resolution and required temporal resolution can be mentioned. Motion estimation information includes, for example, ME (motion estimation) block size, motion prediction mode (intra (in-frame predictive coding) mode, forward predictive coding mode, reverse predictive coding mode, or bidirectional predictive coding mode). , Motion vector, scene change, etc. The time decomposition information includes, for example, information such as the decomposition filter to be used, the time decomposition level of the processing target, and the GOP size.

The reference frame buffer 180 temporarily stores a frame that has been deblocked by the loop filter 170 as a reference frame used for motion estimation by the motion estimation unit 120.

Next, the deblocking filter processing executed by the loop filter 170 of the video coding apparatus 100 having the above configuration will be described. FIG. 2 is a flow diagram for explaining the operation of the deblocking filter processing executed by the loop filter, and the deblocking filtering processing described here is performed for each frame to be processed for each time decomposition level ( For example, it is executed (in order from the first time decomposition level to the final time decomposition level). For example, it is executed for the low frequency frame for each time decomposition level.

The deblock filtering process is started from the acquisition of the ME block in step S1000. Here, one of the ME blocks that make up one time frame is selected and acquired.

That is, in this step, the ME blocks used for motion estimation / compensation are acquired one by one. By doing so, the filter size for deblocking filter processing (tap length described later) can be adapted to the variable block size of the ME block. FIG. 3 is a diagram showing an application example of deblocking filtering based on variable block size. It can be seen that the ME block of the variable block size shown by the solid line in FIG. 3 is matched with the filter size of the deblock filtering process shown by the broken line.

Then, in step S1100, whether or not the other block is adjacent to at least one of the upper end and the left end of the ME block, that is, whether or not there is at least one of the horizontal boundary of the upper end and the vertical boundary of the left end. To judge. In this determination, if there is the above boundary (S1100: YES), the process proceeds to step S1200, and if there is no above boundary (S1100: NO), the process proceeds to step S2200.

In step S1200, one of the acquired ME block boundaries is selected and acquired. Then, in step S1300, the above-mentioned coding / transmission information is acquired.

Then, in step S1400, the tap length of the deblock filtering process applied to the boundary between the ME block and the adjacent block for noise removal is set. The tap length is determined by adopting the smaller of the sizes of two adjacent blocks. Also, in the case of deblocking filtering applied to horizontal boundaries (vertical filtering), the tap length is determined depending on the height of the block, and deblocking filtering applied to vertical boundaries (vertical filtering). In the case of horizontal filtering), the tap length is determined depending on the width of the block.

4 and 5 are diagrams for explaining an example of setting the tap length in the deblocking filter processing, respectively.

In the example shown in FIG. 4, the block P to be processed, the block R adjacent to the upper side of the block P, and the block Q adjacent to the left side of the block P all have the same size. In this case, the tap length of the horizontal filtering applied to the block P is determined based on the heights of the blocks P and Q, and the tap length of the vertical filtering applied to the block P is the block P. , Determined based on the width of R.

On the other hand, in the example shown in FIG. 5, the size of the block P to be processed is smaller than the block Q adjacent to the left side of the block P and larger than each of the blocks R, S, and T adjacent to the upper side of the block P. In this case, the tap length of the horizontal filtering applied to the block P is determined based on the height of the block P. Further, the tap length of the vertical filtering applied to the block P corresponding to the boundary with the block R is determined based on the width of the block R. In addition, the tap length of the vertical filtering applied to the block P corresponding to the boundary with the block S is determined based on the width of the block S, and the tap length of the vertical filtering applied to the block P is determined. The tap length of the one corresponding to the boundary with the block T is determined based on the width of the block T.

In this way, the larger the area where noise is generated, the larger the tap length can be set without waste.

Then, in step S1500, the filter strength when applying the deblock filter processing is set. The filter intensity is set according to the motion prediction mode of two adjacent blocks so that a stronger filter intensity deblock filter process is applied to the blocks with higher noise intensity. For example, when four levels of filter strength can be set, the filter strength is set as follows. When the motion prediction mode of both blocks or one of the blocks is intra, the filter intensity is set to the strongest (Bs = 3). When both blocks refer to different reference frames, when both blocks refer to different numbers of reference frames, or when both blocks refer to the same reference frame, the motion vectors are similar. If not, use the second largest filter strength (Bs = 2). When both blocks refer to the same reference frame and the motion vectors are similar, the third largest (second smallest) filter intensity (Bs = 1) is adopted. In other cases, set the filter strength to off (Bs = 0) so that filtering is not applied to the corresponding boundaries.

Then, in step S1600, the number of pixels to which the deblock filter processing is applied is set. More specifically, the number of applicable pixels in the horizontal filter processing applied to the vertical boundary is set by determining how many pixels to apply the horizontal filter processing from the vertical boundary to the left and right respectively. Further, the number of applicable pixels in the vertical filter processing applied in the horizontal direction is set by determining how many pixels above and below the horizontal boundary the vertical filter processing is applied.

In the case of a vertical boundary, the number of applicable pixels on the left and right sides of the boundary may be set to different values, and in the case of a horizontal boundary, the number of applicable pixels on the upper and lower sides of the boundary may be set to different values. .. However, from the viewpoint of improving processing efficiency and processing speed, it is more effective to set the number of applicable pixels on each of the left and right sides or the number of applicable pixels on each of the upper and lower sides to the same value.

Thresholds are used to determine which pixels to apply filtering to and how many pixels to apply filtering to. This threshold corresponds to the amount of filter required to correct the block noise mixed in due to the lack of layered data for scalability during the coding process and the transmission process. Further, the threshold value may be set empirically according to the coding method used.

In addition, the threshold value is set based on the quantization parameter and the time decomposition level of the time frame applied to the filter. Different quantization parameters generate block noise with different characteristics and magnitudes. If the quantization parameters are not specified in the coding scheme used, the threshold may be derived from the required bit rate or the number of bit planes removed from the bitstream. For example, if the bit rate is low or the number of bit planes removed is large, the quantization parameter will be large. Due to pixel normalization during the MCTF period, the dynamic range of pixel values changes with each time resolution level, but the threshold may be determined by this dynamic range.

Then, in steps S1700, S1800, and S1900, the pixels on the entire line at the acquired boundary are filtered based on the set tap length, filter strength, and number of applied pixels. More specifically, in step S1700, the pixels for one line are filtered, and in step S1800, whether or not the filtering for the final line is completed (whether or not the remaining lines exist at the acquired boundary) is determined. judge. In this determination, if the filtering for the final line is not completed (S1800: NO), the process proceeds to the next line in step S1900 and returns to step S1700. On the other hand, when the filtering for the final line is completed (S1800: YES), the process proceeds to step S2000.

In step S2000, it is determined whether or not there is an unfiltered boundary in addition to the previously acquired boundary in the acquired ME block. In this determination, if another boundary remains (S2000: YES), step S2100 advances to the next boundary and returns to step S1200. On the other hand, if no other boundary remains (S2000: NO), the process proceeds to step S2200. By making such a determination, it is possible to apply filtering to all blocks in the motion estimation / compensation processing. Therefore, this deblocking filter processing can be applied to various reconstruction frames including a time frame at each time decomposition level.

In step S2200, whether or not there are ME blocks that have not yet been filtered in addition to the ME blocks acquired earlier in the time frame currently being processed, in other words, all ME blocks have been filtered. Judge whether or not. In this determination, if another ME block remains (S2200: NO), step S2300 advances to the next ME block and returns to step S1000. On the other hand, if no other ME block remains (S2200: YES), the deblock filter processing at the current time decomposition level is terminated.

By performing the above-mentioned deblocking filter processing, it is possible to perform motion estimation at a subsequent frame or the next time decomposition level by using a cleaner reference frame.

In the deblocking filter processing, in addition to the above steps, the deblocking filter processing execution may be automatically switched on and off according to the acquired color component information.

As already mentioned, the deblocking filter processing performed by the loop filter 170 corresponds to each time decomposition level. That is, as shown in FIG. 6, if the eight original frames at level 0 are time-decomposed into each time frame from level 1 to level 3, then at each of level 1, level 2 and level 3, the above-mentioned de. Block filtering is performed.

As described above, according to the present embodiment, the boundary between a certain motion estimation block in the frame subjected to the motion estimation process and another motion estimation block adjacent to the motion estimation block according to the variable size motion estimation block. Since the deblocking filter processing is applied only by itself, the filter size of the deblocking filtering processing and the motion estimation block size can be matched, and not only the increase in the amount of the deblocking filtering processing can be suppressed, but also the image It is possible to prevent the sharpness of the image from being lost unnecessarily, and it is possible to realize high image quality without consuming unnecessary power of the processing device.

(Embodiment 2) FIG. 7 is a block diagram showing a configuration of a video decoding apparatus having a loop filter according to the second embodiment of the present invention.

In the present embodiment, a general case where a filter for executing the deblock filter processing according to the present invention is used on both the encoder side and the decoder side will be described. Since the filter on the encoder side is the same as the loop filter 170 described in the first embodiment, the description thereof will be omitted.

The video decoding device 200 shown in FIG. 7 includes a reverse scanning / reverse entropy coding unit 210 that performs reverse scanning and reverse entropy coding on a stream input from the corresponding video coding device to generate a frame, and a generation. Spatial wavelet synthesizer 220 that performs spatial wavelet composition on the frame, time filter 230 that performs time filter processing on frames other than independent encoded frames, and time filter processed (time wavelet composition) frame. A motion compensation unit 240 that compensates for motion and a motion-compensated frame are added to generate a reconstruction frame, or a reconstruction frame is generated from an independent encoded frame, and the generated reconstruction frame is generated. The output image addition unit 250, the loop filter 260 that executes the same processing as the deblock filter processing executed by the loop filter 170 described in the first embodiment, and the loop filter 260 perform the deblock filter processing. It has a reference frame buffer 270 that temporarily stores the frame as a reference frame used for motion compensation by the motion compensation unit 240.

The broken line arrow B in FIG. 7 suggests that processing according to a plurality of time decomposition levels is performed in the video decoding apparatus 200.

The loop filter 260 performs the same processing as the deblocking filter processing described in detail in the first embodiment by acquiring the coding / transmission information for adaptively executing the deblocking filter processing separately from the stream. It becomes possible to execute, and thus a cleaner reference frame can be used for time wavelet synthesis by the time filter 230 and motion compensation by the motion compensator 240.

Further, the loop filter 260 receives the signaling from the corresponding video coding device separately from the stream, so that the time decomposition level to which the deblock filter processing is applied is set to a single time decomposition level according to the signaling. Can be adaptively changed from to multiple time decomposition levels. Therefore, when a predetermined instruction is transmitted from the corresponding video coding device, the time decomposition level and the number of times to apply the deblock filter processing can be reduced, and the processing efficiency of the video decoding device 200 is improved. Or the processing load can be reduced.

As described above, according to the present embodiment, the same action and effect as the loop filter in the video coding apparatus described in the first embodiment can be realized by the loop filter in the video decoding apparatus. Further, in order to separately acquire the coding / transmission information from the stream from the corresponding video coding device, the deblocking filter processing executed by each of the video coding device and the video decoding device is the same. This allows the loop filters in each device to operate as a pair of combinations.

In the present embodiment, the coding / transmission information is described as being acquired from the stream from the video coding device, but the video decoding device 200 may derive the coding / transmission information by itself. This is not the case if possible.

(Embodiment 3) FIG. 8 is a block diagram showing a configuration of a video decoding apparatus having a post filter according to the third embodiment of the present invention. The video decoding device of the present embodiment has the same basic configuration as the video decoding device 200 described in the second embodiment, and the same components are designated by the same reference numerals. , The detailed description thereof will be omitted.

In the present embodiment, a general case where the filter that executes the deblocking filter processing according to the present invention is used only on the decoder side will be described.

The video decoding device 300 shown in FIG. 8 has a configuration in which a post filter 310 is provided instead of the loop filter 260 in the video decoding device 200 shown in FIG. 7.

The post filter 310 applies the same processing as the deblocking filter processing described in detail in the first embodiment to the reconstructed frame from the image addition unit 250, and re-creates a cleaner frame to which the deblocking filter processing has been performed. Output as a construction frame.

Further, the post filter 310 separately acquires the coding / transmission information for adaptively executing the deblocking filter processing from the input stream, thereby performing the deblocking filter processing described in detail in the first embodiment. It is possible to execute the same processing as above. However, if the video decoding device 300 can derive the coded / transmitted information by itself, it is not always necessary to acquire the coded / transmitted information from the stream.

Further, the deblock filter process executed by the post filter 310 executes the deblock filter process executed by the loop filter 260 described in the second embodiment at the final level when the reconstruction frame is generated. It is realized by doing.

As described above, according to the present embodiment, the same action and effect as the loop filter in the video coding apparatus described in the first embodiment can be realized by the post filter in the video decoding apparatus.

The deblock filter processing apparatus and the deblock filter processing method of the present invention have the effect of achieving high image quality without consuming unnecessary power of the processing apparatus, and are highly multimedia data coding, particularly variable block. It is useful in video coding using size-based motion estimation.

<figref num="1">The block diagram which shows the structure of the image coding apparatus which has the loop filter which concerns on Embodiment 1 of this invention.</figref><figref num="2">A flow chart for explaining the operation of the deblocking filter processing executed by the loop filter according to the first embodiment of the present invention.</figref><figref num="3">The figure which showed the application example of the variable block size-based deblocking filter processing which concerns on Embodiment 1 of this invention.</figref><figref num="4">The figure for demonstrating an example of tap length setting in the deblocking filter processing which concerns on Embodiment 1 of this invention.</figref><figref num="5">The figure for demonstrating another example of tap length setting in the deblocking filter processing which concerns on Embodiment 1 of this invention.</figref><figref num="6">The figure which shows the relationship between the deblocking filter processing and the time decomposition level which concerns on Embodiment 1 of this invention.</figref><figref num="7">The block diagram which shows the structure of the image decoding apparatus which has the loop filter which concerns on Embodiment 2 of this invention.</figref><figref num="8">The block diagram which shows the structure of the image decoding apparatus which has the post filter which concerns on Embodiment 3 of this invention.</figref><figref num="9">The figure which shows the general example of the conventional MC wavelet coding</figref><figref num="10">Diagram exemplifying variable block size for conventional motion estimation / compensation</figref><figref num="11">A diagram showing an application example of conventional fixed block size-based deblock filtering.</figref>

Code description

100 Video coding device 110 Image input unit 120 Motion estimation unit 130, 230 Time filter 140 Spatial wavelet decomposition unit 150 Scanning / entropy encoding unit 160 Local decoding unit 170, 260 Loop filter 180, 270 Reference frame buffer 200, 300 Video Decoding device 210 Reverse scanning / reverse entropy coding part 220 Spatial wavelet synthesis part 240 Motion compensation part 250 Image addition part 310 Post filter