CN104683819A - Adaptive loop filter method and device - Google Patents
Adaptive loop filter method and device Download PDFInfo
- Publication number
- CN104683819A CN104683819A CN201510052376.1A CN201510052376A CN104683819A CN 104683819 A CN104683819 A CN 104683819A CN 201510052376 A CN201510052376 A CN 201510052376A CN 104683819 A CN104683819 A CN 104683819A
- Authority
- CN
- China
- Prior art keywords
- filter
- buffer memory
- filter factor
- coefficients
- factor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention particularly relates to an adaptive loop filter method and device. According to the method, encoded filter coefficients are cached for using the correlation from frame to frame, brightness components of an image are taken as a whole to train a set of filter coefficients when the filter coefficients are trained, and then the filter coefficients are compared with cached time domain coefficients by a rate distortion optimization method, and a set of optimal filter coefficients are selected to filter the current image; a chrominance component filtering shape is simplified, and the chrominance component filtering shape is 5*5 cross plus 3*3 square. Compared with the prior art, the adaptive loop filter method and device have the beneficial effects of reducing the computational complexity and the hardware cache requirement, saving the resource, and meeting the need of practical application.
Description
Technical field
The present invention is specifically related to a kind of adaptive loop filter method and device, belongs to digital signal processing technique field.
Background technology
Auto-adaptive loop filter (Adaptive Loop-in Filter, ALF) is one of important technology of coding and decoding video.ALF have employed Weiner filter, processes the reconstructed number signal after coding, reduces the mean square error (Mean Square Error, MSE) between primary signal and distorted signal.To encoding, the reconstruction signal obtained carries out filtering, can improve the quality of reconstruction signal, improves the quality of subsequent reference images.
ALF can be divided into three classes according to FILTER TO CONTROL mode: the ALF based on whole frame, the ALF based on region and the ALF based on quad-tree partition.ALF based on whole frame can have larger mass loss for some image block usually, and comparatively complicated based on the ALF dividing mode of quad-tree partition, and encoder complexity is large.Therefore normally used is at present ALF based on region.Entire image is divided into the subregions such as fixing by ALF based on region usually, uses same filter coefficient in same area.Adopt class Hilbert method can for the filter coefficient of selection the best of different region adaptivities, but number of regions is more, spatial cache takies larger, and scan mode is more complicated, and software and hardware implementation complexity is higher.
There is relativity of time domain between adjacent image in usual video, utilize the correlation between image, the image of encoding and decoding during encoding and decoding present image, can be utilized to predict.But existing ALF technology does not make full use of the correlation between image, and the filter factor between different images is relatively independent.In order to reduce temporal redundancy, the method that ALF can adopt time domain filter coefficients multiplexing is to current codec image filtering.
Summary of the invention
In order to simplified calculation method, reduce the requirement of software and hardware, the invention provides a kind of adaptive loop filter method, comprise the steps,
Utilize the correlation between image, buffer memory N overlaps time domain filter coefficients, and the filter factor of buffer memory can be used to carry out filtering to present image;
According to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor; Wherein chromatic component and luminance component adopt different filter shape;
The method of percent of pass aberration optimizing, calculates the rate distortion costs of the filter factor of current training and the time domain filter coefficients of buffer memory, chooses the filter factor of the minimum filter factor of rate distortion costs as present image.
Further, utilize the correlation between image, buffer memory N overlaps time domain filter coefficients, the filter factor of buffer memory can be used to carry out filtering to present image, also comprise, and carries out buffer memory to the filter factor that the image of encoded/decoding uses; Wherein N value is natural number.
Further, wherein chromatic component and luminance component adopt different filter shape, also comprise, and luminance component and chromatic component region sample standard deviation do not carry out Region dividing, only solve a set of filter factor.
Further, according to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor, also comprise luminance channel and chrominance channel, it is square that the filter shape of described luminance channel is that 7x7 cross adds 3x3, has nine coefficients; It is square that the filter shape of described chrominance channel is that 5x5 cross adds 3x3, has seven coefficients.
Further, when buffer memory N cover time domain filter coefficients is not 0, respective filter coefficient comparing rate distortion cost in the luminance channel filter factor that current training is obtained and chrominance channel filter factor and buffer memory, the filter factor as present image that selection rate distortion cost is minimum.
Further, when buffer memory N cover time domain filter coefficients equals 0, namely, when filter factor method is the method without time domain filter coefficients is multiplexing, the luminance channel filter factor directly using current training to obtain and chrominance channel filter factor are as the filter factor of present image.
A kind of adaptive loop filter device, comprises,
Buffer memory time domain filter coefficients unit, for utilizing the correlation between image, buffer memory N frame time domain filter coefficients, can use the filter factor of buffer memory to carry out filtering to present image.
Solve filter coefficients unit, according to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor; Wherein chromatic component and luminance component adopt different filter shape.
Filter factor selected cell, for the method for percent of pass aberration optimizing, calculates the rate distortion costs of the filter factor of current training and the time domain filter coefficients of buffer memory, chooses the filter factor of the minimum filter factor of rate distortion costs as present image.
Further, described buffer memory time domain filter coefficients unit also comprises, and carries out buffer memory for the filter factor used the image of encoded/decoding; Wherein N value is natural number.
Further, described in solve filter coefficients unit and also comprise, luminance component and chromatic component region sample standard deviation do not carry out Region dividing, only solve a set of filter factor.
Further, described in solve filter coefficients unit and also comprise luminance channel and chrominance channel, it is square that the filter shape of described luminance channel is that 7x7 cross adds 3x3, has nine coefficients; It is square that the filter shape of described chrominance channel is that 5x5 cross adds 3x3, has seven coefficients.
Further, when buffer memory N cover time domain filter coefficients is not 0, respective filter coefficient comparing rate distortion cost in the luminance channel filter factor that current training is obtained and chrominance channel filter factor and buffer memory, the filter factor as present image that selection rate distortion cost is minimum.
Further, when buffer memory N cover time domain filter coefficients equals 0, namely, when filter factor method is the method without time domain filter coefficients is multiplexing, the luminance channel filter factor directly using current training to obtain and chrominance channel filter factor are as the filter factor of present image.
The beneficial effect that the present invention is compared to the prior art had: reduce computational complexity and hardware cache demand, saved resource, the demand of realistic application.
Accompanying drawing explanation
Fig. 1 is adaptive loop filter method flow schematic diagram of the present invention;
Fig. 2 is adaptive loop filter device unit figure of the present invention;
Fig. 3 is AVS2ALF Region dividing and the scan mode schematic diagram of adaptive loop filter method of the present invention and device;
Fig. 4 is the ALF filter luminance component filter shape schematic diagram of adaptive loop filter method of the present invention and device;
Fig. 5 is the ALF filter chromatic component filter shape schematic diagram of adaptive loop filter method of the present invention and device;
Fig. 6 is the rate distortion curve synoptic diagram of the cycle tests BQSquare_416x240 of adaptive loop filter method of the present invention and device;
Fig. 7 is the rate distortion curve synoptic diagram of the cycle tests City_1080x720 of adaptive loop filter method of the present invention and device.
Embodiment
When considered in conjunction with the accompanying drawings, by referring to detailed description below, more completely can understand the present invention better and easily learn wherein many adjoint advantages, but accompanying drawing described herein is used to provide a further understanding of the present invention, forms a part of the present invention.
For enabling above-mentioned purpose of the present invention, feature and advantage become apparent more, and the present invention is further detailed explanation with embodiment by reference to the accompanying drawings.
Embodiment: as shown in Fig. 1, Fig. 3 to Fig. 5, present embodiments provides a kind of adaptive loop filter method, comprises the steps,
Step S10, utilizes the correlation between image, and buffer memory N overlaps time domain filter coefficients, and the filter factor of buffer memory can be used to carry out filtering to present image;
Step S20, according to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor; Wherein chromatic component and luminance component adopt different filter shape;
Step S30, the method for percent of pass aberration optimizing, calculates the rate distortion costs of the filter factor of current training and the time domain filter coefficients of buffer memory, chooses the filter factor of the minimum filter factor of rate distortion costs as present image.
In preferred scheme, utilize the correlation between image, buffer memory N overlaps time domain filter coefficients, the filter factor of buffer memory can be used to carry out filtering to present image, also comprise, and carries out buffer memory to the filter factor that the image of encoded/decoding uses; Wherein N value is natural number.
In preferred scheme, wherein chromatic component and luminance component adopt different filter shape, also comprise, and luminance component and chromatic component region sample standard deviation do not carry out Region dividing, only solve a set of filter factor.
In preferred scheme, according to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor, also comprise luminance channel and chrominance channel, it is square that the filter shape of described luminance channel is that 7x7 cross adds 3x3, has nine coefficients; It is square that the filter shape of described chrominance channel is that 5x5 cross adds 3x3, has seven coefficients.
In preferred scheme, when buffer memory N cover time domain filter coefficients is not 0, respective filter coefficient comparing rate distortion cost in the luminance channel filter factor that current training is obtained and chrominance channel filter factor and buffer memory, the filter factor as present image that selection rate distortion cost is minimum.
In preferred scheme, when buffer memory N cover time domain filter coefficients equals 0, namely, when filter factor method is the method without time domain filter coefficients is multiplexing, the luminance channel filter factor directly using current training to obtain and chrominance channel filter factor are as the filter factor of present image.
As shown in Fig. 2, Fig. 3 to Fig. 5, the present embodiment also provides a kind of adaptive loop filter device, comprises,
Buffer memory time domain filter coefficients unit 10, for utilizing the correlation between image, buffer memory N frame time domain filter coefficients, can use the filter factor of buffer memory to carry out filtering to present image.
Solve filter coefficients unit 20, according to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor; Wherein chromatic component and luminance component adopt different filter shape.
Filter factor selected cell 30, for the method for percent of pass aberration optimizing, calculates the rate distortion costs of the filter factor of current training and the time domain filter coefficients of buffer memory, chooses the filter factor of the minimum filter factor of rate distortion costs as present image.
In preferred scheme, described buffer memory time domain filter coefficients unit 10 also comprises, and carries out buffer memory for the filter factor used the image of encoded/decoding; Wherein N value is natural number.
In preferred scheme, described in solve filter coefficients unit 20 and also comprise, luminance component and chromatic component region sample standard deviation do not carry out Region dividing, only solve a set of filter factor.
In preferred scheme, described in solve filter coefficients unit 20 and also comprise luminance channel and chrominance channel, it is square that the filter shape of described luminance channel is that 7x7 cross adds 3x3, has nine coefficients; It is square that the filter shape of described chrominance channel is that 5x5 cross adds 3x3, has seven coefficients.
In preferred scheme, when buffer memory N cover time domain filter coefficients is not 0, respective filter coefficient comparing rate distortion cost in the luminance channel filter factor that current training is obtained and chrominance channel filter factor and buffer memory, the filter factor as present image that selection rate distortion cost is minimum.
In preferred scheme, when buffer memory N cover time domain filter coefficients equals 0, namely, when filter factor method is the method without time domain filter coefficients is multiplexing, the luminance channel filter factor directly using current training to obtain and chrominance channel filter factor are as the filter factor of present image.
Operation principle:
Step S10, utilizes the correlation between image, and buffer memory N overlaps time domain filter coefficients, and the filter factor of buffer memory can be used to carry out filtering to present image;
Step S20, according to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor; Wherein chromatic component and luminance component adopt different filter shape;
Step S30, the method for percent of pass aberration optimizing, calculates the rate distortion costs of the filter factor of current training and the time domain filter coefficients of buffer memory, chooses the filter factor of the minimum filter factor of rate distortion costs as present image.
The technique effect that the present embodiment has is, reduces computational complexity and hardware cache demand, has saved resource, the demand of realistic application.
Table 1, table 2, table 3, Fig. 6 and Fig. 7 list the present invention and realize result on AVS2 reference software RD7.0.Fig. 6 and Fig. 7 lists the distortion performance curve of BQSquare and City sequence when not doing ALF filtering and do ALF filtering respectively.The test performance of table 1, table 2 table 3 and the rate distortion curve of Fig. 6, Fig. 7 show that the ALF filtering method that the present invention proposes significantly can promote coding efficiency.
| Y | U | V | |
| Taishan_1920x1080 | -2.3％ | -2.3％ | -1.7％ |
| City_1080x720 | -3.1％ | -2.9％ | 1.9％ |
| BasketballDrill | -0.9％ | -1.7％ | -2.4％ |
| BQSquare | -0.2％ | 0.0％ | -0.3％ |
| Avg. | -1.6％ | -1.8％ | -1.6％ |
Do ALF filtering and do not do ALF Performance comparision under AI leads to survey condition on table 1.RD7.0 platform
| Y | U | V | |
| Taishan_1920x1080 | -4.14％ | 4.26％ | -3.22％ |
| City_1080x720 | -6.20％ | -4.51％ | -2.99％ |
| BasketballDrill | -0.54％ | -0.55％ | -1.36％ |
| BQSquare | -5.03％ | 0.40％ | 0.23％ |
| Avg. | -4.0％ | -2.2％ | -1.8％ |
Do ALF filtering and do not do ALF Performance comparision under RA leads to survey condition on table 2.RD7.0 platform
| Y | U | V | |
| Taishan_1920x1080 | -6.28％ | -12.06％ | -9.44％ |
| City_1080x720 | -7.84％ | -7.39％ | -7.60％ |
| BasketballDrill | -1.81％ | -0.76％ | -2.41％ |
| BQSquare | -1.46％ | 0.14％ | -3.11％ |
| Avg. | -4.3％ | -5.0％ | -5.6％ |
Do ALF filtering and do not do ALF Performance comparision under LD leads to survey condition on table 3.RD7.0 platform
Be described in detail a kind of adaptive loop filter method provided by the present invention and device above, the exemplary embodiment of above reference accompanying drawing to the application is described.Those skilled in the art should understand that; above-mentioned embodiment is only used to the object that illustrates and the example of lifting; instead of be used for limiting; the any amendment done under all instructions in the application and claims, equivalently to replace, all should be included in and this application claims in the scope of protection.
Claims (12)
1. an adaptive loop filter method, is characterized in that comprising the steps,
Utilize the correlation between image, buffer memory N overlaps time domain filter coefficients, and the filter factor of buffer memory can be used to carry out filtering to present image;
According to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor; Wherein chromatic component and luminance component adopt different filter shape;
The method of percent of pass aberration optimizing, calculates the rate distortion costs of the filter factor of current training and the time domain filter coefficients of buffer memory, chooses the filter factor of the minimum filter factor of rate distortion costs as present image.
2. adaptive loop filter method according to claim 1, it is characterized in that, utilize the correlation between image, buffer memory N overlaps time domain filter coefficients, the filter factor of buffer memory can be used to carry out filtering to present image, also comprise, buffer memory is carried out to the filter factor that the image of encoded/decoding uses; Wherein N value is natural number.
3. adaptive loop filter method according to claim 1, it is characterized in that, wherein chromatic component and luminance component adopt different filter shape, also comprise, luminance component and chromatic component region sample standard deviation do not carry out Region dividing, only solve a set of filter factor.
4. adaptive loop filter method according to claim 1, it is characterized in that, according to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor, also comprise luminance channel and chrominance channel, it is square that the filter shape of described luminance channel is that 7x7 cross adds 3x3, has nine coefficients; It is square that the filter shape of described chrominance channel is that 5x5 cross adds 3x3, has seven coefficients.
5. adaptive loop filter method according to claim 2, it is characterized in that, when buffer memory N cover time domain filter coefficients is not 0, respective filter coefficient comparing rate distortion cost in the luminance channel filter factor that current training is obtained and chrominance channel filter factor and buffer memory, the filter factor as present image that selection rate distortion cost is minimum.
6. adaptive loop filter method according to claim 2, is characterized in that,
When buffer memory N overlap time domain filter coefficients equal 0 time, when namely filter factor method is the method without time domain filter coefficients is multiplexing, directly the luminance channel filter factor that obtains of the current training of use and chrominance channel filter factor are as the filter factor of present image.
7. an adaptive loop filter device, is characterized in that comprising,
Buffer memory time domain filter coefficients unit, for utilizing the correlation between image, buffer memory N frame time domain filter coefficients, can use the filter factor of buffer memory to carry out filtering to present image.
Solve filter coefficients unit, according to compression reconfiguration image and its be compressed image time distortion minimum be target, solve filter factor; Wherein chromatic component and luminance component adopt different filter shape.
Filter factor selected cell, for the method for percent of pass aberration optimizing, calculates the rate distortion costs of the filter factor of current training and the time domain filter coefficients of buffer memory, chooses the filter factor of the minimum filter factor of rate distortion costs as present image.
8. adaptive loop filter device according to claim 7, is characterized in that, described buffer memory time domain filter coefficients unit also comprises, and carries out buffer memory for the filter factor used the image of encoded/decoding; Wherein N value is natural number.
9. adaptive loop filter device according to claim 7, is characterized in that, described in solve filter coefficients unit and also comprise, luminance component and chromatic component region sample standard deviation do not carry out Region dividing, only solve a set of filter factor.
10. adaptive loop filter device according to claim 7, is characterized in that, described in solve filter coefficients unit and also comprise luminance channel and chrominance channel, it is square that the filter shape of described luminance channel is that 7x7 cross adds 3x3, has nine coefficients; It is square that the filter shape of described chrominance channel is that 5x5 cross adds 3x3, has seven coefficients.
11. adaptive loop filter devices according to claim 8, is characterized in that,
When to overlap time domain filter coefficients be not 0 to buffer memory N, respective filter coefficient comparing rate distortion cost in the luminance channel filter factor that current training is obtained and chrominance channel filter factor and buffer memory, the filter factor as present image that selection rate distortion cost is minimum.
12. adaptive loop filter devices according to claim 8, is characterized in that,
When buffer memory N overlap time domain filter coefficients equal 0 time, when namely filter factor method is the method without time domain filter coefficients is multiplexing, directly the luminance channel filter factor that obtains of the current training of use and chrominance channel filter factor are as the filter factor of present image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510052376.1A CN104683819A (en) | 2015-01-31 | 2015-01-31 | Adaptive loop filter method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510052376.1A CN104683819A (en) | 2015-01-31 | 2015-01-31 | Adaptive loop filter method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104683819A true CN104683819A (en) | 2015-06-03 |
Family
ID=53318299
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510052376.1A Pending CN104683819A (en) | 2015-01-31 | 2015-01-31 | Adaptive loop filter method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104683819A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107197291A (en) * | 2017-05-25 | 2017-09-22 | 厦门大学 | The cube convolution Spline Interpolation Method that low complex degree is directly calculated |
| WO2018170801A1 (en) * | 2017-03-22 | 2018-09-27 | 华为技术有限公司 | Image filtering method and apparatus |
| CN109479130A (en) * | 2016-08-02 | 2019-03-15 | 高通股份有限公司 | Adaptive loop filter based on geometric transformation |
| CN109862374A (en) * | 2019-01-07 | 2019-06-07 | 北京大学 | A kind of adaptive loop filter method and device |
| CN112514401A (en) * | 2020-04-09 | 2021-03-16 | 北京大学 | Method and device for loop filtering |
| WO2021129690A1 (en) * | 2019-12-23 | 2021-07-01 | Huawei Technologies Co., Ltd. | Cross-component adaptive loop filtering for video coding |
| WO2021134706A1 (en) * | 2019-12-31 | 2021-07-08 | 北京大学 | Loop filtering method and device |
| CN113785574A (en) * | 2019-05-30 | 2021-12-10 | 北京字节跳动网络技术有限公司 | Adaptive loop filtering of chrominance components |
| WO2022042550A1 (en) * | 2020-08-24 | 2022-03-03 | 杭州海康威视数字技术股份有限公司 | Filtering method and apparatus, and device |
| CN114387192A (en) * | 2021-12-22 | 2022-04-22 | 广东中星电子有限公司 | Image filtering method, image filtering device, electronic equipment and computer readable medium |
| CN114793284A (en) * | 2021-01-26 | 2022-07-26 | 北京达佳互联信息技术有限公司 | System and method for applying adaptive loop filter in video coding and decoding |
| US11490124B2 (en) | 2019-04-20 | 2022-11-01 | Beijing Bytedance Network Technology Co., Ltd. | Signaling of chroma and luma syntax elements in video coding |
| US12041231B2 (en) | 2022-06-29 | 2024-07-16 | SZ DJI Technology Co., Ltd. | In-loop filtering method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120147967A1 (en) * | 2010-12-10 | 2012-06-14 | Qualcomm Incorporated | Adaptive support for interpolating values of sub-pixels for video coding |
| CN102511165A (en) * | 2009-07-31 | 2012-06-20 | 索尼公司 | Image processing apparatus and method |
| CN102714716A (en) * | 2009-08-19 | 2012-10-03 | 索尼公司 | Device and method for image processing |
| CN102857751A (en) * | 2011-07-01 | 2013-01-02 | 华为技术有限公司 | Video encoding and decoding methods and device |
| CN103503456A (en) * | 2011-05-10 | 2014-01-08 | 联发科技股份有限公司 | Method and apparatus for reduction of in-loop filter buffer |
| CN103546746A (en) * | 2009-07-31 | 2014-01-29 | 索尼公司 | Image decoding apparatus, image decoding method, TV receiver and portable telephone |
-
2015
- 2015-01-31 CN CN201510052376.1A patent/CN104683819A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102511165A (en) * | 2009-07-31 | 2012-06-20 | 索尼公司 | Image processing apparatus and method |
| CN103546746A (en) * | 2009-07-31 | 2014-01-29 | 索尼公司 | Image decoding apparatus, image decoding method, TV receiver and portable telephone |
| CN102714716A (en) * | 2009-08-19 | 2012-10-03 | 索尼公司 | Device and method for image processing |
| US20120147967A1 (en) * | 2010-12-10 | 2012-06-14 | Qualcomm Incorporated | Adaptive support for interpolating values of sub-pixels for video coding |
| CN103503456A (en) * | 2011-05-10 | 2014-01-08 | 联发科技股份有限公司 | Method and apparatus for reduction of in-loop filter buffer |
| CN102857751A (en) * | 2011-07-01 | 2013-01-02 | 华为技术有限公司 | Video encoding and decoding methods and device |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11095922B2 (en) | 2016-08-02 | 2021-08-17 | Qualcomm Incorporated | Geometry transformation-based adaptive loop filtering |
| CN109479130A (en) * | 2016-08-02 | 2019-03-15 | 高通股份有限公司 | Adaptive loop filter based on geometric transformation |
| CN109479130B (en) * | 2016-08-02 | 2021-04-20 | 高通股份有限公司 | Adaptive loop filtering based on geometric transformation |
| WO2018170801A1 (en) * | 2017-03-22 | 2018-09-27 | 华为技术有限公司 | Image filtering method and apparatus |
| US11190778B2 (en) | 2017-03-22 | 2021-11-30 | Huawei Technologies Co., Ltd. | Image filtering method and apparatus |
| CN107197291B (en) * | 2017-05-25 | 2020-02-14 | 厦门大学 | Low-complexity direct-calculation cubic convolution spline interpolation method |
| CN107197291A (en) * | 2017-05-25 | 2017-09-22 | 厦门大学 | The cube convolution Spline Interpolation Method that low complex degree is directly calculated |
| CN109862374A (en) * | 2019-01-07 | 2019-06-07 | 北京大学 | A kind of adaptive loop filter method and device |
| US11490124B2 (en) | 2019-04-20 | 2022-11-01 | Beijing Bytedance Network Technology Co., Ltd. | Signaling of chroma and luma syntax elements in video coding |
| US12022130B2 (en) | 2019-04-20 | 2024-06-25 | Beijing Bytedance Network Technology Co., Ltd. | Signaling of syntax elements for joint coding of chrominance residuals |
| US11575939B2 (en) | 2019-04-20 | 2023-02-07 | Beijing Bytedance Network Technology Co., Ltd. | Signaling of syntax elements for joint coding of chrominance residuals |
| CN113785574B (en) * | 2019-05-30 | 2022-10-11 | 北京字节跳动网络技术有限公司 | Adaptive loop filtering of chrominance components |
| CN113785574A (en) * | 2019-05-30 | 2021-12-10 | 北京字节跳动网络技术有限公司 | Adaptive loop filtering of chrominance components |
| US11477449B2 (en) | 2019-05-30 | 2022-10-18 | Beijing Bytedance Network Technology Co., Ltd. | Adaptive loop filtering for chroma components |
| WO2021129690A1 (en) * | 2019-12-23 | 2021-07-01 | Huawei Technologies Co., Ltd. | Cross-component adaptive loop filtering for video coding |
| WO2021134706A1 (en) * | 2019-12-31 | 2021-07-08 | 北京大学 | Loop filtering method and device |
| CN112514401A (en) * | 2020-04-09 | 2021-03-16 | 北京大学 | Method and device for loop filtering |
| WO2021203394A1 (en) * | 2020-04-09 | 2021-10-14 | 北京大学 | Loop filtering method and apparatus |
| WO2022042550A1 (en) * | 2020-08-24 | 2022-03-03 | 杭州海康威视数字技术股份有限公司 | Filtering method and apparatus, and device |
| CN114793284A (en) * | 2021-01-26 | 2022-07-26 | 北京达佳互联信息技术有限公司 | System and method for applying adaptive loop filter in video coding and decoding |
| CN114793284B (en) * | 2021-01-26 | 2023-06-27 | 北京达佳互联信息技术有限公司 | System and method for applying adaptive loop filter in video codec |
| CN114387192A (en) * | 2021-12-22 | 2022-04-22 | 广东中星电子有限公司 | Image filtering method, image filtering device, electronic equipment and computer readable medium |
| CN114387192B (en) * | 2021-12-22 | 2024-05-03 | 广东中星电子有限公司 | Image filtering method, device, electronic equipment and computer readable medium |
| US12041231B2 (en) | 2022-06-29 | 2024-07-16 | SZ DJI Technology Co., Ltd. | In-loop filtering method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104683819A (en) | Adaptive loop filter method and device | |
| CN104702963A (en) | Boundary treatment method and device for adaptive loop filter | |
| TW201526617A (en) | Method and system for image processing, decoding method, encoder and decoder | |
| JP2017523668A (en) | System and method for high content adaptive quality restoration filtering for video coding | |
| CN104735450A (en) | Method and device for carrying out adaptive loop filtering during video coding and decoding | |
| CN104584561A (en) | Method and apparatus for sample adaptive offset in a video decoder | |
| KR20210132708A (en) | Loop filtering implementation method, apparatus and computer storage medium | |
| KR20220036982A (en) | Quadratic transformation for video encoding and decoding | |
| EP3706046A1 (en) | Method and device for picture encoding and decoding | |
| WO2020009758A1 (en) | Context-based binary arithmetic encoding and decoding | |
| CN105306957A (en) | Adaptive loop filtering method and device | |
| US20240187568A1 (en) | Virtual temporal affine candidates | |
| CN105245896A (en) | HEVC (High Efficiency Video Coding) parallel motion compensation method and device | |
| US11750836B2 (en) | Virtual temporal affine candidates | |
| CN104521234A (en) | Method and apparatus for video processing incorporating deblocking and sample adaptive offset | |
| US20220124337A1 (en) | Harmonization of intra transform coding and wide angle intra prediction | |
| US20220417516A1 (en) | Deep intra prediction of an image block | |
| JP2022537173A (en) | Method and device for picture encoding and decoding using position-dependent intra-prediction combination | |
| KR20220123643A (en) | Sub-block merging candidates in triangle merging mode | |
| CN104486630B (en) | H.265 in Video coding under SAO patterns offset statistical method | |
| EP3611924A1 (en) | In-loop filter with multiple regions | |
| CN112703733A (en) | Translation and affine candidates in a unified list | |
| US20230370645A1 (en) | In-loop filter with multiple regions | |
| CN114554213B (en) | Motion adaptive and detail-focused compressed video quality enhancement method | |
| US10491927B1 (en) | Quadtree-based bandwidth compression prediction method and system thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150603 |