Residual Coefficient Scanning Regions for Lower Bitstream Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards, such as AVS3, inefficiently define the last non-zero coefficient and lack a flag to indicate non-zero coefficients within coding groups, leading to inefficient signaling for every position.
Innovation Solution
The proposed solution involves coding the scanning region of coefficients in a non-rectangular shape, using a signaled scanning order index, and restricting the bottom-right position of the scanning region to be unequal to the block's bottom-right corner, allowing for more efficient coding by indicating non-zero coefficients and adapting the scanning order based on block dimensions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a flag is signaled for every position to indicate non-zero coefficients, then non-zero coefficients can be accurately indicated, but bitstream complexity increases
Solution Approach 1:
The transform block is divided into multiple coding groups (CGs), each containing multiple coefficient positions. Instead of signaling flags for every individual position, a single flag is signaled per coding group to indicate whether any non-zero coefficients exist within that group. This segmentation approach reduces the number of signaled flags from N (one per position) to M (one per group), where M < N, thereby reducing bitstream complexity while maintaining accurate indication of non-zero coefficient regions.
2Productivity
If the scanning region includes the entire current block, then all coefficients can be coded, but the coding efficiency decreases due to unnecessary signaling of zero coefficients
Solution Approach 1:
The patent extracts and identifies the actual scanning region containing non-zero coefficients separately from the entire current block. By using the bottom-right position (brPos) of the scanning region, the method extracts only the necessary portion of the block that contains meaningful data. This allows the coding process to focus exclusively on relevant coefficients while implicitly handling zero coefficients outside the scanning region, thereby improving coding efficiency without losing essential information.
3Area of stationary object
If the bottom-right position of the scanning region is allowed to be the block's bottom-right corner, then the entire block can be scanned, but the ability to indicate non-zero regions efficiently is reduced
Solution Approach 1:
The patent applies local quality by allowing the bottom-right position (brPos) of the scanning region to be dynamically set to different positions within the current block based on the actual distribution of non-zero coefficients. Instead of always using the block's bottom-right corner, the method adapts the scanning region boundary to match the local characteristics of the transform coefficients. This localized adaptation enables efficient indication of non-zero regions by restricting the scanning area to where actual data exists, reducing unnecessary signaling while maintaining complete coverage of relevant coefficients.
Data Source
AI summary
Video processing methods, devices, and systems are provided. In one example aspect, a method of video processing includes performing a conversion between a current block of a video and a coded representation of the video. The current block has a width W and a height H and includes an allowable scanning region that includes non-zero transform coefficients that are coded in the coded representation according to a scanning order. The allowable scanning region is restricted to a portion of the current block according to a rule that specifies that a bottom-right corner position of the allowable scanning region is different than a bottom-right corner position of the current block.