KR-20260066168-A - Adaptive clipping in the method of deriving model parameters for video compression

KR20260066168AKR 20260066168 AKR20260066168 AKR 20260066168AKR-20260066168-A

Abstract

The method includes the step of receiving a bitstream of coded information for a picture. The coded information directs a prediction of the current block using a model-based prediction technique, wherein the model-based prediction technique generates a prediction sample of the current block based on a model having one or more reconstructed samples of a reference block input to the model, and the model includes one or more parameters derived based on a current template of the current block and a reference template of the reference block. The method further includes the step of performing at least one clipping operation on at least one of the current template and the reference template to obtain a clipped template sample; the step of deriving one or more parameter values of one or more parameters of the model according to the clipped template sample; and the step of generating at least one prediction sample of the current block by the model.

Inventors

체르냐크 로만
왕 뱌오
전 리옌-페이
샹 쯔위에
윤용욱
쉬 모퉁
리우 샨

Assignees

텐센트 아메리카 엘엘씨

Dates

Publication Date: 20260512
Application Date: 20241025
Priority Date: 20241022

Claims (15)

As a video decoding method, A step of receiving a bitstream of coded information for a sequence of pictures - said coded information indicates a prediction of a current block within a current picture using a model-based prediction technique, said model-based prediction technique generates a prediction sample of said current block based on a model having one or more reconstructed samples of a reference block input to the model, said model includes one or more parameters derived based on a current template of said current block and a reference template of said reference block -; A step of performing at least one clipping operation on at least one of the current template and the reference template to obtain a clipped template sample; A step of deriving one or more parameter values of one or more parameters of the model according to the clipped template sample; and A step of generating at least a prediction sample of the current block by the model having the one or more parameters set to the one or more parameter values. A video decoding method including
In paragraph 1, A step of performing a clipping operation on a reference sample of the reference block to obtain a clipped reference sample; and Step of applying the model to the clipped reference sample to generate a prediction sample of the current block A video decoding method that further includes
In paragraph 1, A video decoding method further comprising the step of performing a clipping operation on the prediction sample of the current block after the prediction sample is generated by the model.
In paragraph 1, The step of performing at least the clipping operation described above is A step of performing a first clipping operation on the current template to obtain a clipped current template sample; and A step of performing a second clipping operation on the reference template to obtain a clipped reference template sample A video decoding method that further includes
In paragraph 1, A step of checking whether the current template and the reference template are affected by a mapping function; and When both the current template and the reference template are not affected by the mapping function, A step of performing a first clipping operation on the current template to obtain a clipped current template sample, and A step of performing a second clipping operation on the reference template to obtain a clipped reference template sample A video decoding method that further includes
In paragraph 5, A step of checking whether the above mapping function is enabled; When the mapping function is enabled, a step of checking whether the current template and the reference template are affected by the mapping function; and A step of performing the first clipping operation and the second clipping operation when the current template and the reference template are not affected by the mapping function. A video decoding method that further includes
In paragraph 1, A step of checking whether the current template and/or the reference template has a value domain different from the original picture of the current picture; and When the templates of the current template and the reference template have different value domains from the original picture, a step of performing a clipping operation on the template to obtain a clipped template sample of the template A video decoding method that further includes
In Paragraph 7, The step of performing a clipping operation on the above template is: A video decoding method comprising the step of applying the clipping operation to the template after the inverse mapping function is applied to the template.
In Paragraph 7, The step of performing a clipping operation on the above template is: A video decoding method comprising the step of applying the clipping operation to the template before the inverse mapping function is applied to the template.
In Paragraph 9, A video decoding method further comprising the step of calculating a clipping range for the clipping operation according to a mapping function associated with the inverse mapping function.
In Paragraph 7, The step of performing a clipping operation on the above template is: A video decoding method comprising the step of applying the clipping operation to the template without applying an inverse mapping function to the template before and after the clipping operation.
In paragraph 1, A step of checking whether the current template and/or the reference template has the same value domain as the original picture of the current picture; and When the templates of the current template and the reference template have the same value domain as the original picture, a step of performing a clipping operation on the template to obtain a clipped template sample of the template before deriving one or more parameter values. A video decoding method that further includes
In Paragraph 12, A video decoding method comprising the step of applying the clipping operation to the template after the mapping function is applied to the template.
As a video encoding method, A step of determining to code a current block within a current picture using a model-based prediction technique - the model-based prediction technique generates a prediction sample of the current block based on a model having one or more reconstructed samples of a reference block input to the model, and the model includes one or more parameters derived based on a current template of the current block and a reference template of the reference block -; A step of performing at least one clipping operation on at least one of the current template and the reference template to obtain a clipped template sample; A step of deriving one or more parameter values of one or more parameters of the model according to the clipped template sample; and A step of encoding the current block based on the model having one or more parameters set to one or more parameter values. A video encoding method including
As a method for processing visual media data, It includes a step of processing a bitstream of visual media data according to format rules, and The bitstream includes coded information of a current block within a current picture, and the coded information indicates the prediction of a current block within a current picture using a model-based prediction technique, and the model-based prediction technique generates a prediction sample of the current block based on a model having one or more reconstructed samples of a reference block input to the model, and the model includes one or more parameters derived based on a current template of the current block and a reference template of the reference block. The above format rules are, At least one clipping operation is performed on at least one of the current template and the reference template to obtain a clipped template sample; One or more parameter values of one or more parameters of the model are derived according to the clipped template sample; and At least one predicted sample of the current block is generated by the model having the one or more parameters set to the one or more parameter values. A method for processing visual media data that specifies

Description

Adaptive clipping in the method of deriving model parameters for video compression Integration by reference This application claims the benefit of priority to U.S. Patent Application No. 18/923,643, filed October 22, 2024, under the title "Adaptive Clipping in Models Parameters Derivations Methods for Video Compression," and this U.S. patent application claims the benefit of priority to U.S. Provisional Application No. 63/602,343, filed November 22, 2023, under the title "Adaptive Clipping in Models Parameters Derivations Methods for Video Compression." The disclosures of the prior applications are hereby incorporated by reference in their entirety. The present disclosure describes aspects generally related to video coding. The background description provided herein is intended for the purpose of generally presenting the context of the disclosure. The work of the inventors currently named, to the extent that the work is described in this background section, as well as aspects of the description that would not otherwise qualify as prior art at the time of filing, are not recognized as prior art with respect to the disclosure, either expressly or impliedly. Image/video compression can help transmit image/video data across different devices, storage, and networks with minimal quality degradation. In some examples, video codec technology can compress video based on spatial and temporal redundancy. In one example, a video codec may utilize a technique referred to as intra prediction, which can compress images based on spatial redundancy. For instance, intra prediction may use reference data from the current picture being reconstructed for sample prediction. In another example, a video codec may utilize a technique referred to as inter prediction, which can compress images based on temporal redundancy. For instance, inter prediction may predict samples within the current picture from a previously reconstructed picture with motion compensation. Motion compensation can be indicated by a motion vector (MV). Aspects of the disclosure include a bitstream, a method, and an apparatus for video encoding/decoding. In some examples, the apparatus for video encoding/decoding includes a processing circuit. Some aspects of the disclosure provide a video decoding method. The method comprises receiving a bitstream of coded information for a sequence of pictures, wherein the coded information directs a prediction of a current block in the current picture using a model-based prediction technique, wherein the model-based prediction technique generates a prediction sample of the current block based on a model having one or more reconstructed samples of a reference block input to the model, and the model includes one or more parameters derived based on a current template of the current block and a reference template of the reference block. The method further comprises: performing at least one clipping operation on at least one of the current template and the reference template to obtain a clipped template sample; deriving one or more parameter values of one or more parameters of the model according to the clipped template sample; and generating at least one prediction sample of the current block by the model having one or more parameters set to one or more parameter values. Some aspects of the disclosure provide a video encoding method. The method comprises the step of determining to code a current block in a current picture using a model-based prediction technique, wherein the model-based prediction technique generates a prediction sample of the current block based on a model having one or more reconstructed samples of a reference block input to the model, and the model includes one or more parameters derived based on a current template of the current block and a reference template of the reference block. The method further comprises the step of performing at least one clipping operation on at least one of the current template and the reference template to obtain a clipped template sample; the step of deriving one or more parameter values of one or more parameters of the model according to the clipped template sample; and the step of encoding the current block based on a model having one or more parameters set to one or more parameter values. A side of the disclosure also provides a device for video encoding/decoding. A part of the disclosure also provides a method for processing visual media data. In the method, a bitstream of visual media data is processed according to a format rule. For example, the bitstream may be a bitstream that is decoded/encoded in any of the decoding and/or encoding methods described herein. The format rule may specify one or more constraints of the bitstream and/or one or more processes to be performed by the decoder and/or encoder. A side of the disclosure also provides a non-transient computer-readable medium storing instructions that, when executed by a computer, cause the computer to perform any of t