Adaptive Color Transform in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges such as inefficient chroma BDPCM mode usage, potential QP becoming negative when ACT is enabled, lack of support for lossless coding in some ACT designs, and suboptimal signaling of ACT usage and palette sizes.
Innovation Solution
The proposed solution involves enabling chroma BDPCM mode based on ACT and luma BDPCM mode usage, clipping the QP when ACT is enabled, supporting lossless coding in ACT designs, and dynamically signaling ACT usage and palette sizes based on coding characteristics and block dimensions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If chroma BDPCM mode is used independently without considering ACT mode, then chroma coding simplicity is maintained, but coding efficiency is reduced due to suboptimal mode selection
Solution Approach 1:
The patent makes the chroma BDPCM mode selection dynamic by linking it to the ACT mode state. When ACT mode is enabled, chroma BDPCM mode is automatically enabled as well, creating an adaptive mode selection mechanism that responds to the coding characteristics of the current block rather than using a fixed selection rule
Solution Approach 2:
The patent implements a feedback mechanism where the state of ACT mode (determined by coding characteristics such as palette mode usage or intra prediction mode) directly influences the selection of chroma BDPCM mode. This feedback loop ensures that chroma residual coding is optimized based on the actual characteristics of the luma component coding
2Adaptability or versatility
If QP is allowed to become negative when ACT is enabled, then coding flexibility is improved, but reliability deteriorates due to potential overflow or underflow issues
Solution Approach 1:
The patent applies preliminary anti-action by pre-clipping the QP value before it can become negative. The clipping operation ensures that QP remains within a valid range (QP ≥ 0) by adjusting any negative QP values to zero, thereby preventing potential overflow or underflow issues that would arise from negative QP values while ACT mode is enabled
3Adaptability or versatility
If ACT design does not support lossless coding, then implementation complexity is reduced, but adaptability deteriorates due to inability to handle lossless coding scenarios
Solution Approach 1:
The patent enables lossless coding support in ACT by introducing parameter changes - specifically, by setting the transform skip flag to indicate that the ACT transform should be skipped during lossless coding scenarios. This allows the ACT framework to adapt to both lossy and lossless coding modes by simply changing the state of the transform skip flag, without requiring fundamentally different implementation paths
4Productivity
If ACT usage and palette sizes are signaled with fixed methods, then signaling simplicity is maintained, but coding efficiency deteriorates due to inability to adapt to different block characteristics
Solution Approach 1:
The patent implements dynamic signaling where the method of signaling ACT usage and palette sizes adapts based on the characteristics of the current video block. For example, palette mode signaling is applied conditionally based on whether the block is suitable for palette coding, and the maximum palette size is determined dynamically rather than being fixed, allowing the signaling to be optimized for each specific block type
Solution Approach 2:
The patent applies local quality by using different signaling strategies for different regions or types of blocks within the video stream. Instead of using a uniform signaling method for all blocks, the patent selectively applies palette mode signaling, ACT usage signaling, and maximum palette size signaling based on the specific characteristics of each block, thereby optimizing the signaling for local coding requirements
Data Source
AI summary
A method for implementing an adaptive colour transform (ACT) mode during image/video encoding and decoding includes performing a conversion between a current video block of a video and a bitstream of the video, wherein the current video block is coded using the ACT mode, wherein the conversion includes applying an inverse ACT transform on the current video block according to a rule, and wherein the rule specifies that a clipping operation, based on a bit depth of the current video block, is applied to an input of the inverse ACT transform.