Video Coding Quantization Group Signaling for Non-Square CUs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently signaling quantization information when using non-square coding units, leading to disruptions in quantization group schemes and reduced compression performance.
Innovation Solution
The proposed solution involves redefining quantization groups as successive coding units or blocks with boundaries aligning with CU boundaries and sizes greater than or equal to a threshold, allowing for efficient signaling of local quantization information and improved compression performance by using QTBT or MTT partitioning schemes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If non-square CUs are used in video coding, then coding flexibility and compression performance are improved, but quantization group scheme is disrupted and signaling complexity increases
Solution Approach 1:
The patent segments the video picture into quantization groups that are independent of CU shape, allowing non-square CUs to be processed while maintaining regular quantization group structures. This segmentation decouples the CU partitioning flexibility from the quantization group signaling complexity.
Solution Approach 2:
The patent applies local quantization parameters specifically at quantization group boundaries rather than uniformly across all CUs. This allows differentiated quantization control for different regions while maintaining a simplified overall quantization group scheme that works with non-square CUs.
2Adaptability or versatility
If non-square CUs are used, then coding adaptability is improved, but signaling of quantization information becomes more complex
Solution Approach 1:
The patent segments quantization information signaling into quantization group-level parameters and CU-level adjustments. This segmentation reduces the overall signaling overhead by grouping quantization information at a higher level while still allowing local adaptations for non-square CUs.
Solution Approach 2:
Instead of signaling quantization information for each CU individually (which would increase overhead with non-square CUs), the patent inverts the approach by signaling quantization group parameters first and then deriving CU-specific quantization information from these group parameters, reducing total signaling overhead.
3Measurement precision
If quantization groups are redefined to align with CU boundaries, then quantization accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent applies enhanced quantization accuracy locally at quantization group boundaries where it is most needed, while using standard processing for other regions. This localized approach improves overall quantization accuracy without uniformly increasing processing complexity across the entire video picture.
Solution Approach 2:
The patent performs preliminary organization of CUs into quantization groups before the actual quantization process. This preliminary action aligns CU boundaries with quantization group boundaries in advance, simplifying the subsequent quantization processing while maintaining high quantization accuracy.
Data Source
AI summary
A video decoder receives local quantization information for a current quantization group. The video decoder determines a partitioning of a coding tree unit (CTU) of a picture of the video data into a plurality of coding units (CUs). Additionally, the video decoder derives, based at least in part on the local quantization information for the current quantization group, a quantization parameter. The current quantization group is defined as a group of successive CUs so that boundaries of the current quantization group must be boundaries of the CUs. The current quantization group may or may not be square. Additionally, the video decoder inverse quantizes, based on the quantization parameter, at least one transform coefficient of a current CU being in the current quantization group. The video decoder reconstructs, based on inverse quantized transform coefficients of the current CU, a coding block of the current CU.