Bi-directional Image Decoding Weight Index Entropy Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image encoding and decoding methods face challenges in reducing residual data size and simplifying entropy encoding and decoding processes, particularly in achieving low bit rates and efficient data compression.
Innovation Solution
The implementation of bi-directional prediction in image encoding and decoding, where weight information from a bitstream is used to combine reference blocks, with entropy decoding of weight indices to reconstruct current blocks, allowing for adaptive selection of candidate values based on Picture Order Count (POC) and accumulative selection frequencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If bi-directional prediction with weight information is used to combine reference blocks, then compression efficiency is improved and bit rate is reduced, but device complexity increases due to additional entropy decoding and combining operations
Solution Approach 1:
The patent performs preliminary actions by preparing multiple candidate weight values and their corresponding binary representations in advance. The encoder determines the optimal weight value before encoding, and the decoder reconstructs it using predetermined combining operations. This preliminary preparation reduces the complexity of real-time decision-making during encoding/decoding while maintaining high compression efficiency through optimal weight selection.
Solution Approach 2:
The patent introduces dynamic adaptability by allowing the weight value to vary based on picture order count (POC) differences and accumulative selection frequencies. The system dynamically adjusts which candidate weight value is selected and how many bits are allocated for encoding, optimizing the balance between compression efficiency and complexity based on current video content characteristics and temporal reference relationships.
2Quantity of substance
If residual data size is reduced through bi-directional prediction, then bit rate is lowered, but the complexity of entropy encoding and decoding processes increases
Solution Approach 1:
The patent segments the weight index encoding into multiple binary values, where each binary value corresponds to a specific bit position. This segmentation allows the entropy encoder to process each bit independently using context models, simplifying the encoding process while achieving efficient compression. The segmented approach transforms a complex multi-value encoding problem into simpler binary decision problems.
Solution Approach 2:
The patent applies partial action by selectively allocating bits for weight index encoding based on the magnitude of POC differences. When POC differences are small, fewer bits are allocated; when differences are large, more bits are allocated. This partial encoding approach optimizes the balance between residual data reduction and encoding complexity by applying entropy encoding only where necessary rather than uniformly across all cases.
3Measurement precision
If adaptive weight selection based on POC and selection frequency is implemented, then prediction accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent implements feedback mechanisms by maintaining accumulative selection frequency counters for each candidate weight value. The system uses this accumulated feedback information to adaptively adjust encoding parameters, such as the number of bits allocated for weight index encoding and the selection of context models. This feedback-driven adaptation improves prediction accuracy by learning from historical performance while managing processing complexity through incremental updates rather than complete re-evaluation.
Solution Approach 2:
The patent changes encoding parameters dynamically based on POC differences and selection frequencies. The number of bits allocated for weight index encoding is adjusted as a parameter based on the magnitude of temporal reference differences. Context model parameters are also changed adaptively based on accumulated selection frequencies. These parameter changes enable the system to optimize prediction accuracy for different video content characteristics without requiring complete re-implementation of the encoding logic.
Data Source
AI summary
Disclosed is an image decoding method according to an embodiment, the image decoding method including: obtaining a first reference block and a second reference block, for bi-directional prediction of a current block; obtaining, from a bitstream, weight information for combining the first reference block with the second reference block; performing entropy decoding on the weight information to obtain a weight index; combining the first reference block with the second reference block according to a candidate value indicated by the weight index among candidate values included in a weight candidate group; and reconstructing the current block based on a result of the combining, wherein a first binary value corresponding to the weight index is entropy-decoded based on a context model, and the remaining binary value corresponding to the weight index is entropy-decoded by a bypass method.