Neural Intra Prediction Coding for Lower Entropy Hardware Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face high hardware implementation complexity due to sequential entropy coding of syntax elements related to intra prediction modes, which requires complex hardware implementations.
Innovation Solution
An encoding and decoding method using a neural network to perform probability estimation on intra prediction mode syntax elements, allowing entropy encoding/decoding to be performed only on an index of a syntax element value set, reducing hardware complexity by leveraging dependency relationships between syntax elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If sequential entropy coding is performed on multiple syntax elements related to intra prediction mode, then complete prediction mode information is encoded, but hardware implementation complexity increases significantly
Solution Approach 1:
The patent combines multiple syntax elements (MIP flag, MIP mode, multi-row reference index, ISP mode, MPM flag, MPM index, and intra prediction mode index) into a single unified syntax element called intra prediction mode syntax element value set. This single syntax element encapsulates all the information that would otherwise require sequential coding of multiple separate elements, thereby reducing hardware complexity while preserving complete prediction mode information.
Solution Approach 2:
The intra prediction mode syntax element value set serves multiple functions simultaneously: it indicates whether MIP mode is used, specifies the MIP mode value, indicates multi-row reference usage, specifies ISP mode, indicates MPM mode, specifies MPM index, and provides the final intra prediction mode index. This multi-functional single element replaces what would traditionally require multiple separate syntax elements and sequential coding operations.
2Loss of information
If multiple syntax elements are coded sequentially, then all prediction mode parameters are transmitted, but encoding time and processing steps increase
Solution Approach 1:
The patent merges the coding operations for multiple syntax elements into a single entropy coding operation on the intra prediction mode syntax element value set. This consolidation reduces the number of sequential coding steps from multiple individual element codings to one unified coding operation, thereby reducing encoding time while ensuring all prediction mode parameters are transmitted.
3Adaptability or versatility
If traditional sequential entropy coding is used for intra prediction mode syntax elements, then standard compatibility is maintained, but compression efficiency improvement is limited
Solution Approach 1:
The patent changes the parameter structure by transforming multiple separate syntax elements into a single unified syntax element value set with an associated index. This parameter transformation enables more efficient entropy coding by allowing the use of probability models that can better exploit the statistical dependencies among the previously separate elements, thereby improving compression efficiency while maintaining standard compatibility through the systematic definition of the new syntax element structure.
Data Source
AI summary
An encoding method, a decoding method, and a device are provided. The encoding method includes: determining an index of an intra prediction mode syntax element value set of a current picture block; performing, by using a neural network, probability estimation on input data that represents a feature of the current picture block, to obtain probability distribution of a plurality of candidate intra prediction mode syntax element value sets, wherein the probability distribution represents respective probability values of the plurality of candidate intra prediction mode syntax element value sets; and performing entropy encoding on a probability value related to the index of the intra prediction mode syntax element value set of the current picture block based on the probability distribution of the plurality of candidate intra prediction mode syntax element value sets, to obtain a bitstream.