Circuit structure and method for realizing loop filtering, electronic component and electronic equipment

CN119031148APending Publication Date: 2024-11-26BEIJING BOYA RUISHI TECH CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411073525.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-06
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing loop filtering method has a long and complex processing time in hardware implementation, resulting in a large amount of hardware resource occupancy and it is difficult to meet the needs of efficient coding.

Method used

By adjusting the SAO parameter decision order, pipeline processing of DB filtering and SAO filtering is realized, and original pixels and reconstructed pixels are uniformly cached to reduce RAM usage. U components and V components share processing circuits, and alternately processed according to pixel sub-blocks.

Benefits of technology

It significantly reduces the filtering processing time, reduces hardware resource usage, improves processing efficiency, and meets the requirements of high real-timeness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119031148A_ABST
    Figure CN119031148A_ABST
Patent Text Reader

Abstract

The invention provides a circuit structure for realizing loop filtering, which comprises an SAO parameter decision module and a fusion filtering module, and is characterized in that the SAO parameter decision module is used for executing an SAO parameter decision according to an original pixel and a reconstructed pixel of a current LCU to obtain a compensation parameter; and the fusion filtering module is used for executing DB filtering and SAO filtering on the reconstructed pixels in a pipeline mode, and applying the compensation parameters in the SAO filtering. According to the technical scheme of the invention, the order of SAO parameter decisions is adjusted, the SAO compensation parameter decisions are subjected to approximate processing, assembly line processing of DB filtering and SAO filtering is realized, the overall processing speed of filtering is accelerated, and a hardware product meeting the requirement of high real-time performance can be obtained.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of video coding technology, and in particular to a circuit structure, method, electronic component and electronic device for implementing loop filtering. Background Art

[0002] In the video coding process, the temporal redundant information in the reconstructed image can be used through motion estimation technology to improve the compression efficiency of the coding. However, the transformation and quantization processing in the coding process will cause loss of the quality of the reconstructed image. In order to improve the quality of the reconstructed image, the loop filtering process is introduced in video coding. The first two levels of filtering in the loop filtering process are DB (Deblocking) filtering and SAO (Sample Adaptive Offset) filtering. DB filtering is used to eliminate the block effect in the reconstructed image, and SAO filtering is used to eliminate the ringing effect of the strong edge in the reconstructed image.

[0003] In the current common loop filtering method, DB filtering and SAO filtering are not executed continuously. The decision process of SAO compensation parameters is inserted in the middle. This decision process requires statistics of pixel information in each maximum coding unit (LCU). Therefore, it is necessary to cache the DB filtering results of each LCU, and SAO filtering can only be performed after the parameter decision is completed. This not only increases the processing time of the entire loop filtering circuit, but also increases the cache area and the complexity of hardware implementation.

[0004] Therefore, how to improve the processing efficiency of loop filtering during hardware implementation is an urgent problem to be solved. Summary of the invention

[0005] The present invention provides a circuit structure, method, electronic component and electronic equipment device for realizing loop filtering, which are used to solve the problems of long processing time and complex hardware realization in the existing loop filtering method.

[0006] To solve the above technical problems, the present invention is achieved as follows:

[0007] In one aspect, the present disclosure provides a circuit structure for implementing loop filtering, comprising:

[0008] SAO parameter decision module and fusion filter module, where:

[0009] The SAO parameter decision module is used to perform SAO parameter decision according to the original pixels and reconstructed pixels of the current LCU to obtain compensation parameters;

[0010] The fusion filtering module is used to perform DB filtering and SAO filtering on the reconstructed pixels in a pipeline manner, and apply the compensation parameters in the SAO filtering.

[0011] In another aspect, the present disclosure further provides a method for implementing loop filtering, comprising:

[0012] Perform SAO parameter decision according to the original pixels and reconstructed pixels of the current LCU to obtain decision parameters;

[0013] DB filtering and SAO filtering are performed on the reconstructed pixels in a pipeline manner, and the compensation parameters are applied in the SAO filtering.

[0014] On the other hand, the present disclosure further provides an electronic component for video encoding, which includes the aforementioned circuit architecture for implementing loop filtering.

[0015] On the other hand, the present disclosure further provides an electronic device, which includes the electronic component for video encoding as described above, so that the electronic device can encode the video image.

[0016] According to the technical solution disclosed in the present invention, by adjusting the order of SAO parameter decision, the decision of SAO compensation parameters is approximated, and the pipeline processing of DB filtering and SAO filtering is realized, which speeds up the overall processing speed of filtering. It has been verified by experiments that the BD-Rate (Bjontegaard Delta-Rat, an indicator for evaluating the rate-distortion performance of video encoders) loss of the approximate processing of SAO parameter decision is less than 0.15%, and the filtering processing time is reduced by nearly half. On this basis, the original pixels, reconstructed pixels, and DB filtered pixels are uniformly cached, and the scattered small-size RAMs can be merged into a unified large-size RAM, thereby saving hardware area. In the process of SAO parameter decision, DB filtering, and SAO filtering, the U component and the V component share a processing circuit, and are processed alternately and serially according to pixel sub-blocks, which further reduces the occupation of hardware resources, thereby helping to obtain hardware products that meet high real-time requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a schematic diagram of a traditional loop filter structure;

[0019] Figure 2 A circuit architecture block diagram of a loop filter according to an embodiment of the present disclosure;

[0020] Figure 3A schematic diagram of DB filtering pixels at a vertical boundary in AVS3 in an embodiment of the present disclosure;

[0021] Figure 4 Schematic diagram of the internal pipeline processing process of DB filtering and SAO filtering in the embodiment of the present disclosure;

[0022] Figure 5 Schematic diagram of the YUV processing structure in the SAO parameter decision module in the embodiment of the present disclosure;

[0023] Figure 6 is a schematic diagram of a YUV processing structure in a fusion filter module in an embodiment of the present disclosure;

[0024] Figure 7 A schematic diagram of the structure of a loop filter circuit provided by another embodiment of the present disclosure;

[0025] Figure 8 A schematic diagram of a flow chart of a method for implementing loop filtering provided in an embodiment of the present disclosure;

[0026] Fig. 9 A flowchart of a method for implementing loop filtering provided by another embodiment of the present disclosure is provided. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the technical solutions in the present disclosure, the technical solutions of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present disclosure. In addition, for the sake of clarity, parts that are not related to the description of the exemplary embodiments are omitted in the drawings.

[0028] In the present disclosure, it should be understood that terms such as "including" or "having" are intended to indicate the existence of features, numbers, steps, behaviors, components, parts, or combinations thereof disclosed in the present disclosure, and are not intended to exclude the possibility that one or more other features, numbers, steps, behaviors, components, parts, or combinations thereof exist or are added. It should also be noted that the embodiments in the present disclosure and the features in the embodiments may be combined with each other if there is no conflict.

[0029] The circuit structure and method for implementing loop filtering provided by the present disclosure can be applied to standard or non-standard image or video encoders. For example, encoders that comply with VVC, HEVC, and AVS3 standards. In order to clearly illustrate the specific implementation of the technical solution of the present disclosure, the AVS3 coding standard is used as an example in the following description.

[0030] Before describing the circuit structure and method for implementing loop filtering disclosed in the present invention, the conventional loop filtering method currently used is first introduced.

[0031] Figure 1 A schematic diagram of the traditional loop filter structure is given.

[0032] like Figure 1 As shown in the figure, the original pixels and reconstructed pixels of the LCU are input and cached in different RAMs. The reconstructed pixels are first subjected to DB filtering to obtain DB filtered pixels, which are then cached. The SAO parameter decision is made based on the DB filtered pixels and the original pixels. The obtained decision parameters are used as an input of the SAO filter. The SAO filter reads the DB filtered pixels from the cache and then performs SAO filtering on them. Finally, the filtered pixels are output.

[0033] exist Figure 1 In the execution order of the traditional filtering shown, DB filtering and SAO filtering cannot be executed continuously, and the SAO parameter decision process needs to be completed in the middle. Therefore, in the hardware implementation process, the DB filter pixels need to be cached at the LCU level before SAO filtering can be performed. For example, in AVS3, the maximum size of the LCU supports 128×128, and the RAM for caching DB filter pixels needs to be increased accordingly, which will significantly increase the hardware area.

[0034] Therefore, the technical solution provided by the present disclosure adjusts the processing order of "DB filtering-SAO parameter decision-SAO filtering" in the traditional method, gives priority to SAO parameter decision, and realizes pipeline processing after fusing DB filtering with SAO filtering. On this basis, by using pixel sub-blocks smaller than LCU as processing units, the original pixels, reconstructed pixels, and DB filtered pixels with similar cache reading and writing methods are uniformly cached and managed, which reduces the implementation complexity, reduces the LCU-level cache of DB filtered pixels, and further reduces the hardware area growth caused by different RAM sizes.

[0035] The implementation methods of various technical solutions of the present disclosure are described in detail below with reference to the accompanying drawings.

[0036] Figure 2 The circuit architecture block diagram of the loop filter according to the embodiment of the present disclosure is shown in FIG.

[0037] like Figure 2 As shown, the circuit architecture 100 includes a SAO parameter decision module 110 and a fusion filter module 120 .

[0038] The SAO parameter decision module 110 is used to perform SAO parameter decision according to the original pixels and reconstructed pixels of the current LCU to obtain compensation parameters.

[0039] The fusion filtering module 120 is used to perform DB filtering and SAO filtering on the reconstructed pixels in a pipeline manner, and apply the compensation parameters in the SAO filtering.

[0040] According to the embodiment of the present disclosure, the SAO parameter decision module uses the original pixel and the reconstructed pixel before DB filtering for difference statistics, while the traditional method uses the original pixel and the pixel after DB filtering, but the performance loss caused by this difference is extremely small. According to experimental tests, the BD-Rate loss caused by the approximate processing here using the technical solution of the present disclosure is less than 0.15%, and the processing time reduction brought by pipeline processing through fusion filtering is close to half of the original method.

[0041] Preferably, the fusion filter module divides the reconstructed pixels of the current LCU into multiple n×n pixel sub-blocks, and performs DB filtering and SAO filtering on each pixel sub-block in a pipeline manner. The use of pixel sub-blocks is conducive to pipeline hardware implementation. The fusion filter module specifically includes a DB filter module and an SAO filter module. The DB filter module is used to perform DB filtering on the pixel sub-block to obtain a DB filtered pixel block, and the SAO filter module is used to perform SAO filtering on the DB filtered pixel block. The speed at which the DB filter module processes the pixel sub-block is consistent with the speed at which the SAO filter module processes the DB filtered pixel block, so that a more efficient pipeline processing effect can be obtained.

[0042] When the DB filter module completes the filtering of the first pixel sub-block, it continues to process the second pixel sub-block. At the same time, the SAO filter module processes the first DB filter pixel block. When the SAO filter module completes the filtering of the first DB filter pixel block, the DB filter module also completes the filtering of the third pixel sub-block. The filtering process of subsequent sub-blocks is similar. When the DB filter module completes processing all pixel sub-blocks of the current LCU, the SAO filter module processes the last DB filter pixel block. Therefore, the overall filtering time is the sum of the DB filter time for one LCU plus the SAO filter time for one DB filter pixel sub-block. In the traditional method, the overall filtering time is the sum of the DB filter time for one LCU plus the SAO filter time for one LCU. In the technical solution disclosed in the present invention, the fusion filter utilizes the existing DB filter module and SAO filter module, but due to the use of a pipeline processing structure, the processing efficiency is greatly improved.

[0043] The value of n is a value that can divide the LCU into an integer number of sub-blocks, such as 4, 8, 16, 32, etc., and can be selected according to the boundary division method and pixel offset method defined in the coding standard. For example, in the AVS3 standard, for DB filtering, it is necessary to filter the division boundary, specifically involving vertical boundaries and horizontal boundaries, such as Figure 3As shown in the figure, the dotted line represents the vertical boundary, the 4 pixels on the left side of the boundary are recorded as L0~L3, and the 4 pixels on the right side of the boundary are recorded as R0~R3. In the DB filtering process, 4 pixels on the left and right sides of the boundary are read, and the filtering strength is calculated according to the DB filtering related parameters. According to the filtering strength, at most 3 filtered pixels on the left and right sides are written back, namely L0~L2 and R0~R2. For the horizontal boundary, the filtering process is the same. Since the reconstructed pixels of the right LCU or the bottom LCU will be used when the DB filtering is processed to the boundary of the LCU, in order to reduce the dependency, in the hardware implementation, the processing range of the DB filter is offset 4 pixels to the left and upward for processing. The offset processing method introduces 4 rows of image-level pixel caches and 4 columns of LCU-level pixel caches to save the unfiltered pixels on the right and bottom sides of the current LCU after the offset. In AVS3 encoding, SAO filtering is also performed based on a range of 4 pixels offset to the left and upward. Therefore, DB filtering and SAO filtering can be combined for synchronous processing, and the size of the pixel sub-block can be set to 8×8.

[0044] In this case, if the number of clocks for processing an 8×8 sub-block is 8, for a 64×64 LCU, the number of clocks for filtering in the traditional way is the sum of the DB filtering clock number and the SAO filtering clock number, that is:

[0045] ((64×64) / (8×8))×8+((64×64) / (8×8))×8=1024

[0046] In the disclosed embodiment, the number of filtering clocks in the pipeline mode after the DB filter and the SAO filter are integrated is: ((64×64) / (8×8))×8+8=520. Compared with the traditional filtering mode, the filtering speed is nearly doubled.

[0047] According to this embodiment, a schematic diagram of the internal pipeline processing process of DB filtering and SAO filtering is as follows: Figure 4 As shown, the processing unit for each filtering is an 8×8 pixel block, and the processing speed of the DB filter module for each processing unit is synchronized with the processing speed of the SAO filter module for each processing unit. After the DB filter module completes the filtering of the first 8×8 processing unit, it outputs the filtered pixels to the SAO filter module. When the SAO filter module processes the first 8×8 processing unit, the DB filter module processes the second 8×8 processing unit. The filtering process of the subsequent 8×8 processing units is carried out in sequence until the filtering process of all pixels in an LCU is completed.

[0048] Figure 5 It is a schematic diagram of the YUV processing structure in the SAO parameter decision module.

[0049] like Figure 5As shown, the SAO parameter decision module in the circuit structure of this embodiment includes a Y component statistical module and a UV component statistical module in parallel with each other, and a YUV parameter decision module. Among them, the Y component statistical module is used to count the Y component of the original pixel of the current LCU and the Y component of the reconstructed pixel of the current LCU; the UV component statistical module is used to count the UV component of the original pixel of the current LCU and the UV component of the reconstructed pixel of the current LCU. The U component and the V component share a statistical circuit, and the statistics of the U component and the V component are performed alternately according to the pixel sub-block. Each statistical module outputs the difference between the original pixel and the reconstructed pixel under different types, as well as the number of pixels of the corresponding type. The YUV parameter decision module is used to perform parameter decision based on the statistical results of the Y component statistical module and the UV component statistical module to obtain compensation parameters. Specifically, the YUV parameter decision module decides the optimal compensation parameter from the sideband mode, boundary mode, and fusion mode according to the statistical difference and number, and the compensation parameter is used in the subsequent SAO filtering link.

[0050] Figure 6 It is a schematic diagram of the YUV processing structure in the fusion filter module.

[0051] like Figure 6 As shown, the fusion filter module includes a parallel Y component filter branch and a UV component filter branch. The Y component filter branch is used to perform DB filtering and SAO filtering on the Y component in the reconstructed pixels of the current LCU; the UV component filter branch is used to perform DB filtering and SAO filtering on the UV component in the reconstructed pixels of the current LCU, and alternately perform filtering on the U component and the V component according to the pixel sub-block, that is, the processing process of the Y component is parallel to the processing process of the UV component, and the UV component shares one processing circuit, and is processed alternately in series according to the U component and the V component. Figure 6 The filtering structure in each branch is as follows: Figure 4 As shown, the Y component filtering branch or the UV component filtering branch includes a DB filtering module and an SAO filtering module, wherein the DB filtering module is used to perform DB filtering on the pixel sub-block to obtain a DB filtered pixel block, and the SAO filtering module is used to perform SAO filtering on the DB filtered pixel block according to the compensation parameter, and the speed at which the DB filtering module processes the pixel sub-block is consistent with the speed at which the SAO filtering module processes the DB filtered pixel block.

[0052] Figure 7 A schematic diagram of the structure of a loop filter circuit provided in another embodiment of the present disclosure.

[0053] In the boundary mode, the SAO filter needs the DB filter pixels around the LCU as reference pixels. Therefore, when the DB filter reaches the boundary of the LCU, the filter result needs to be cached so as to serve as the SAO filter reference pixel of the right LCU or the bottom LCU. At the same time, in combination with the cache link in the filtering process, the circuit structure of the disclosed embodiment also includes a unified cache module. The unified cache module is used to cache the original pixels and reconstructed pixels of the current LCU, as well as some reconstructed pixels obtained after DB filtering to obtain DB filter pixel blocks, that is, DB filter pixel blocks at the LCU boundary. After processing in the above manner, the LCU-level cache of the DB filter can be reduced, and the unified pixel cache module caches the original pixels, reconstructed pixels and filtered pixels with similar read and write methods by splicing, which can reduce the amount of RAM, thereby further reducing the area occupied by the RAM.

[0054] like Figure 7 As shown, the circuit structure 200 of this embodiment includes an SAO parameter decision module 210, a fusion filter module 220, and a unified cache module 230. The original pixels and reconstructed pixels of the LCU are input to the unified cache module 230, and the SAO parameter decision module 210 reads the original pixels and reconstructed pixels therefrom. The internal circuit is divided into two statistical circuits, Y and UV, for calculation (such as Figure 5 As shown). The statistical process of the Y component is processed in parallel with the statistical process of the UV component, and a statistical circuit is shared inside the UV component. The UV component is counted alternately by pixel sub-block. The statistical module outputs the difference between the original pixel and the reconstructed pixel under different types, as well as the number of pixels of the corresponding type. The parameter decision module 210 decides the optimal compensation parameter from the sideband mode, boundary mode, and fusion mode according to the difference and the number, and the compensation parameter is used as the input of the subsequent SAO filtering. After obtaining the compensation parameter, the fusion filter module 220 performs filtering, that is, performs DB filtering and SAO filtering. During the filtering process, the reconstructed pixels are read from the unified cache module 230, and the DB filtered pixels at the boundary are written into the unified cache module 230 as reference pixels required for subsequent processing.

[0055] The disclosed embodiment places all RAM storage related to pixel cache in a unified cache module for processing. For original pixels, reconstructed pixels, and filtered pixels with similar cache read and write methods, the scattered small-sized RAMs can be merged into large-sized RAMs by splicing to reduce the number of RAMs and further reduce the area occupied by RAMs.

[0056] According to the above embodiments, the technical solution disclosed in the present invention executes the SAO parameter decision process independently in advance, and then merges the DB filter with the SAO filter. On the one hand, it reduces the LCU-level DB filter pixel cache required by the traditional filtering process, and on the other hand, it merges the original serial processing time of the DB filter and the SAO filter into the processing time of only the DB filter, which is nearly twice as fast as the traditional method. On this basis, a unified cache module is introduced to reduce the amount of RAM, further reduce the area occupied by the RAM, and reduce the hardware area growth caused by different RAM sizes. In addition, since the number of pixels of the U component and the V component is only half of the number of pixels of the Y component, the UV component can also share a processing circuit, which does not affect the overall processing speed and can further reduce the hardware logic resource occupation. Compared with the traditional loop filtering method, it has significant beneficial effects.

[0057] Based on the same inventive concept, an embodiment of the present disclosure also provides a loop filtering method.

[0058] Figure 8 It is a flowchart of a method for implementing loop filtering provided by an embodiment of the present disclosure.

[0059] like Figure 8 As shown, the loop filtering method provided by the embodiment of the present disclosure includes operations S310 to S320.

[0060] S310: Execute SAO parameter decision using original pixels and reconstructed pixels of the current LCU to obtain compensation parameters.

[0061] S320: Perform DB filtering and SAO filtering on the reconstructed pixels in a pipeline manner, and use the compensation parameter in the SAO filtering.

[0062] According to the method of this embodiment, S320 specifically includes: dividing the reconstructed pixels of the current LCU into a plurality of n×n pixel sub-blocks, and performing DB filtering and SAO filtering on each pixel sub-block in a pipeline manner.

[0063] According to this implementation, the method further includes, before S310, storing original pixels and reconstructed pixels of the current LCU in a unified cache; in S320, storing DB filtered pixels obtained by DB filtering some reconstructed pixels in the unified cache.

[0064] According to this implementation, S310 specifically includes:

[0065] According to the original pixels and reconstructed pixels of the current LCU, Y component statistics and UV component statistics are performed in parallel, wherein the Y component statistics are performed based on the Y component of the original pixel and the Y component of the reconstructed pixel, and the UV component statistics are performed based on the UV component of the original pixel and the UV component of the reconstructed pixel. The UV component statistics alternately perform statistics of the U component and the V component according to the pixel sub-block;

[0066] Based on the Y component statistics and UV component statistics, parameter decision is performed to obtain compensation parameters.

[0067] Furthermore, S320 may also include:

[0068] DB filtering and SAO filtering are performed on the Y component in the reconstructed pixel; in parallel, DB filtering and SAO filtering are performed on the UV component in the reconstructed pixel, and filtering of the U component and the V component is performed alternately on a pixel sub-block basis.

[0069] Combined with the above processing flow, Fig. 9 A flow chart of a loop filtering method according to another embodiment of the present disclosure is given. Fig. 9 As shown, this embodiment includes operations S410 to S450.

[0070] S410: The unified cache module receives original pixels and reconstructed pixels.

[0071] S420: Execute SAO parameter decision.

[0072] S430: According to the decision result, DB filtering and SAO filtering are processed in a pipeline.

[0073] S440: The unified cache module updates the cache DB filtered pixels.

[0074] S450: Output filtering results.

[0075] The specific implementation of the method of this embodiment can refer to the description in the above embodiment, which will not be repeated here. For subsequent processing, the SAO filtering result is also output to the reference pixel cache as the reference pixel of the subsequent inter-frame prediction module, and the SAO decision result is also output to the entropy coding module. These processes are not within the scope of this disclosure and will not be repeated here.

[0076] According to the technical solution disclosed in the present invention, by adjusting the order of SAO parameter decision, the decision of SAO compensation parameters is approximated, and the pipeline processing of DB filtering and SAO filtering is realized, which speeds up the overall processing speed of filtering. It has been verified by experiments that the BD-Rate loss of SAO parameter decision approximation is less than 0.15%, and the filtering processing time is reduced by nearly half. On this basis, the original pixels, reconstructed pixels, and DB filtered pixels are uniformly cached, which can reduce the hardware area growth caused by different RAM sizes, and in the process of SAO parameter decision, DB filtering and SAO filtering, the U component and the V component share a processing circuit, and are processed alternately and serially according to pixel sub-blocks, which further reduces the occupation of hardware resources, thereby helping to obtain hardware products that meet high real-time requirements.

[0077] On this basis, the embodiment of the present disclosure further provides an electronic component for video encoding, which includes the aforementioned circuit architecture for implementing loop filtering, so that the electronic component can perform various loop filtering methods as described above. The electronic component can be used to implement an encoder that meets the requirements of ultra-high-definition video.

[0078] On this basis, the embodiment of the present disclosure also provides an electronic device, which includes the electronic components for video encoding as described above, so that the electronic device can encode and process video images. For example, a smart phone, computer or other device including the aforementioned electronic components can realize ultra-high-definition video communication to meet user needs.

[0079] The above describes specific embodiments of the present invention. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0080] Each embodiment in the present disclosure is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the embodiments of electronic components and electronic devices, since they are basically similar to the embodiments of circuit structures and methods, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0081] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0082] The above description is only an embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various changes and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the scope of the claims of the present disclosure.

Claims

1. A circuit structure for implementing loop filtering, characterized in that: include: SAO parameter decision module and fusion filter module, where: The SAO parameter decision module is used to perform SAO parameter decision according to the original pixels and reconstructed pixels of the current LCU to obtain compensation parameters; The fusion filtering module is used to perform DB filtering and SAO filtering on the reconstructed pixels in a pipeline manner, and apply the compensation parameters in the SAO filtering.

2. The circuit structure according to claim 1, characterized in that: The fusion filtering module divides the reconstructed pixels of the current LCU into a plurality of n×n pixel sub-blocks, and performs DB filtering and SAO filtering on each pixel sub-block in a pipeline manner.

3. The circuit structure according to claim 2, characterized in that: The SAO parameter decision module includes a Y component statistics module and a UV component statistics module in parallel, and a YUV parameter decision module, wherein: The Y component statistics module is used to perform statistics on the Y components of the original pixels of the current LCU and the Y components of the reconstructed pixels of the current LCU; The UV component statistics module is used to perform statistics on the UV components of the original pixels of the current LCU and the UV components of the reconstructed pixels of the current LCU, and perform statistics on the U component and the V component alternately according to the pixel sub-block; The YUV parameter decision module is used to perform parameter decision based on the statistical results of the Y component statistical module and the UV component statistical module to obtain the compensation parameters of the current LCU.

4. The circuit structure according to claim 2, characterized in that: The fusion filter module includes a parallel Y component filter branch and a UV component filter branch, wherein: The Y component filtering branch is used to perform DB filtering and SAO filtering on the Y component in the reconstructed pixels of the current LCU; The UV component filtering branch is used to perform DB filtering and SAO filtering on the UV component in the reconstructed pixels of the current LCU, and to perform filtering on the U component and the V component alternately according to the pixel sub-blocks.

5. The circuit structure according to claim 4, characterized in that: The Y component filtering branch or the UV component filtering branch includes a DB filtering module and an SAO filtering module, the DB filtering module is used to perform DB filtering on the pixel sub-block to obtain a DB filtered pixel block, the SAO filtering module is used to perform SAO filtering on the DB filtered pixel block according to the compensation parameter, and the speed at which the DB filtering module processes the pixel sub-block is consistent with the speed at which the SAO filtering module processes the DB filtered pixel block.

6. The circuit structure according to claim 1, characterized in that: The circuit structure further includes a unified cache module, which is used to cache the original pixels and reconstructed pixels of the current LCU, and DB filtered pixels obtained after DB filtering of part of the reconstructed pixels.

7. A method for implementing loop filtering, characterized in that: include: Perform SAO parameter decision according to the original pixels and reconstructed pixels of the current LCU to obtain decision parameters; DB filtering and SAO filtering are performed on the reconstructed pixels in a pipeline manner, and the compensation parameters are applied in the SAO filtering.

8. The method according to claim 7, characterized in that The performing DB filtering and SAO filtering on the reconstructed pixels in a pipeline manner includes: The reconstructed pixels of the current LCU are divided into a plurality of n×n pixel sub-blocks, and DB filtering and SAO filtering are performed on each pixel sub-block in a pipeline manner.

9. The method according to claim 7, characterized in that: The method further comprises: Before performing SAO parameter decision according to the original pixels and reconstructed pixels of the current LCU to obtain compensation parameters, the original pixels and reconstructed pixels of the current LCU are stored in a unified buffer; A DB filtered pixel block is obtained by DB filtering part of the pixel sub-blocks and stored in the unified cache.

10. The method according to claim 1, characterized in that The performing SAO parameter decision according to the original pixel and the reconstructed pixel of the current LCU to obtain the compensation parameter includes: According to the original pixels and the reconstructed pixels of the current LCU, Y component statistics and UV component statistics are performed in parallel, wherein the Y component statistics are performed based on the Y component of the original pixel and the Y component of the reconstructed pixel, the UV component statistics are performed based on the UV component of the original pixel and the UV component of the reconstructed pixel, and the UV component statistics alternately perform statistics of the U component and the V component according to the pixel sub-block; Based on the Y component statistics and the UV component statistics, parameter decision is performed to obtain compensation parameters.

11. The method according to claim 8, characterized in that The performing DB filtering and SAO filtering on the reconstructed pixels in a pipeline manner includes: DB filtering and SAO filtering are performed on the Y component in the reconstructed pixel; in parallel, DB filtering and SAO filtering are performed on the UV component in the reconstructed pixel, and filtering of the U component and the V component is performed alternately according to the pixel sub-block.

12. An electronic component for video encoding, characterized in that: The invention comprises a circuit structure for realizing loop filtering as claimed in any one of claims 1 to 6.

13. An electronic device, characterized in that: The electronic device comprises the electronic component for video encoding as claimed in claim 12, so that the electronic device can perform encoding processing on the video image.

Citation Information

Cited By

  • Pixel adaptive compensation method and system, electronic equipment, medium and program product

    CN120512537A