Adaptive Transform Scanning in Video Decoding Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video signal compression methods face inefficiencies due to inadequate or excessive elimination of spatial and temporal redundancies, leading to reduced compression rates or degraded decoding performance.
Innovation Solution
Adaptive definition of unit block sizes for transform processes, use of flag information to indicate block sizes and types, application of suitable transforms like DCT or KLT based on prediction modes, and adaptive scanning orders to optimize coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If spatial redundancy and temporal redundancy are excessively eliminated to improve compression rate, then compression rate is improved, but decoding performance is degraded due to loss of information required for decoding
Solution Approach 1:
The patent applies parameter changes by adaptively selecting transform block sizes (4x4, 8x8, 16x16) and transform types (DST, DCT, KLT) based on prediction modes and block characteristics. This allows optimization of compression rate while preserving necessary decoding information through appropriate transform selection rather than excessive redundancy elimination.
Solution Approach 2:
The patent implements dynamics through adaptive transform processing where the transform block size and type are dynamically selected based on the prediction mode and block characteristics. The system transitions from static transform application to dynamic adaptation, improving compression efficiency while maintaining decoding quality by selecting optimal transforms for each block.
2Device complexity
If a fixed transform block size is used to simplify the decoding process, then device complexity is reduced, but coding efficiency is lowered due to inability to adapt to different block characteristics
Solution Approach 1:
The patent applies segmentation by dividing the macroblock into different transform block sizes (4x4, 8x8, 16x16) based on the prediction mode and block characteristics. This segmentation allows the decoder to handle different block sizes efficiently while improving coding efficiency through adaptive transform selection for each segment.
Solution Approach 2:
The patent implements local quality by applying different transform types (DST, DCT, KLT) to different transform blocks based on their specific characteristics and prediction modes. Each block receives a locally optimized transform rather than a uniform approach, improving overall coding efficiency while maintaining manageable decoding complexity.
Data Source
AI summary
The present invention relates to a method for decoding a video signal, comprising the steps of: acquiring a transform size flag of the current macroblock from a video signal; checking the number of non-zero transform coefficients at each pixel position in a first transform block which corresponds to the transform size flag; changing a scan order of the first transform block by prioritizing the position of the pixel having the greatest number of the nonzero transform coefficients in the first transform block; determining the number of the non-zero transform coefficients at each pixel position in a second transform block, and setting the changed scan order of the first transform block as an initialized scan order of the second transform block; adding the number of the non-zero transform coefficients at each pixel position in the first transform block and the number of the non-zero transform coefficients at each pixel position in the second transform block, and changing the scan order of the second transform block by prioritizing the position of the pixel having the greatest number of the nonzero transform coefficients; and decoding the transform coefficients arranged in the scan order changed in the previous step, wherein the first transform block and the second transform block have sizes corresponding to the transform size flag, and are contained in the current macroblock.