Intra TMP Template Selection for Accurate Low-Complexity Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding technologies face issues of large deviation and low prediction accuracy in Intra Template Matching Prediction (Intra TMP), leading to reduced compression efficiency and performance.
Innovation Solution
A method that combines multiple candidate templates and block vectors for Intra TMP, allowing the encoder and decoder to determine a block vector candidate list and prediction value based on selected templates, enhancing compression efficiency and decoding performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single template is used for Intra TMP, then the decoding complexity is reduced, but the prediction accuracy deteriorates
Solution Approach 1:
The patent segments the template selection process by dividing multiple candidate templates into different groups or categories. The decoder receives indication information that points to a specific candidate template from the candidate template list, allowing the system to evaluate multiple templates without requiring the decoder to perform exhaustive template matching operations, thus maintaining low decoding complexity while improving prediction accuracy.
Solution Approach 2:
The patent performs preliminary template selection at the encoder side by constructing a candidate template list and selecting the most suitable candidate template before transmission. The encoder evaluates multiple templates and determines the best match in advance, then transmits only the indication information to the decoder. This preliminary action transfers the computational burden to the encoder, allowing the decoder to achieve high prediction accuracy with minimal complexity.
2Measurement precision
If multiple candidate templates are evaluated at the decoder, then the prediction accuracy is improved, but the decoding complexity increases
Solution Approach 1:
The patent implements a feedback mechanism where the encoder transmits indication information to the decoder about which candidate template from the pre-defined list should be used. This feedback allows the decoder to directly select the appropriate template without performing its own template matching evaluations, thereby maintaining low decoding complexity while achieving high prediction accuracy through the encoder's preliminary evaluation.
Solution Approach 2:
The patent introduces an intermediary element - the candidate template list with indication information - that mediates between the encoder's template evaluation and the decoder's template selection. The indication information acts as a mediator that carries the essential selection decision from the encoder to the decoder, eliminating the need for the decoder to perform complex template matching while ensuring accurate template selection.
3Measurement precision
If the template search range is expanded, then the prediction accuracy is improved, but the computational cost increases
Solution Approach 1:
The patent applies local quality by focusing template matching efforts on the most promising regions. Instead of uniformly evaluating all possible templates in an expanded search range, the system uses the candidate template list that prioritizes templates with higher likelihood of being optimal. This allows the encoder to achieve high prediction accuracy by concentrating computational resources on the most relevant candidates rather than exhaustively searching the entire expanded range.
Solution Approach 2:
The patent changes the parameter of template evaluation by using a candidate template list that is pre-ordered or pre-filtered based on certain criteria. Rather than evaluating templates in a fixed search order across the entire search range, the system transforms the template selection problem into selecting from a curated list of candidates, effectively changing the search strategy parameter to reduce computational cost while maintaining accuracy.
Data Source
AI summary
A method for decoding includes: a first template corresponding to a current block is determined, where the first template includes one or more candidate templates in a template set corresponding to the current block; a block vector candidate list corresponding to the current block is determined based on the first template; and block vector corresponding to the current block is determined based on the block vector candidate list corresponding to the current block, and a prediction value of the current block is determined based on the block vector.