Intra Template Matching with Multi-Template Block Vector Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding 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 candidates with multiple templates in Intra TMP, constructing a block vector candidate list and selecting a first template to improve compression efficiency and decoding performance by enhancing the encoding and decoding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional Intra TMP uses a single template for prediction, then the encoding process is simple, but the prediction accuracy is low and compression efficiency is reduced
Solution Approach 1:
The patent segments the template selection process by dividing templates into multiple groups (first template group and second template group) with different priorities. The encoder first selects from the first template group, and only if no valid template is found does it proceed to the second template group. This segmentation reduces the search complexity while maintaining prediction accuracy by organizing templates hierarchically.
Solution Approach 2:
The patent performs preliminary actions by pre-organizing templates into grouped structures and establishing selection rules before the actual encoding process. The block vector candidate list is constructed in advance with prioritized template groups, allowing the decoder to quickly determine prediction templates without exhaustive search, thus reducing runtime complexity while preserving accuracy.
2Measurement precision
If multiple candidate templates are searched exhaustively, then prediction accuracy improves, but encoding time and computational cost increase
Solution Approach 1:
The patent implements a dynamic template selection mechanism where the search process adapts based on findings. The encoder dynamically switches between template groups - starting with the first template group and only proceeding to the second group if necessary. This dynamic approach reduces average encoding time by avoiding unnecessary searches through the entire template set while maintaining the ability to achieve high prediction accuracy when needed.
Solution Approach 2:
The patent applies partial action by performing template matching only to the extent necessary. Instead of exhaustively evaluating all candidate templates, the encoder performs partial searches within the first template group and only extends to the second group if the first group yields no valid results. This partial evaluation approach significantly reduces encoding time while maintaining sufficient prediction accuracy for most cases.
3Device complexity
If a limited template set is used, then encoding complexity is reduced, but prediction accuracy and compression efficiency deteriorate
Solution Approach 1:
The patent makes the template selection system universal by designing a multi-functional template group structure that can adapt to different coding scenarios. The first template group handles common prediction cases efficiently, while the second template group provides additional options for difficult cases. This universal structure allows the same encoding framework to achieve both low complexity for simple cases and high compression efficiency for complex cases without requiring separate systems.
Solution Approach 2:
The patent employs a nested structure where the second template group serves as a backup layer within the overall template selection system. The template groups are nested in terms of selection priority - the first template group is the primary layer, and the second template group is the secondary layer that is only accessed when the first layer fails to provide a valid template. This nesting allows the system to maintain low complexity for most operations while having access to an expanded template set when needed for improved compression efficiency.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Disclosed in the embodiments of the present application are an encoding/decoding method, a code stream, an encoder, a decoder and a storage medium. The method comprises: an encoder/decoder determining a first template corresponding to the current block, wherein the first template is one or more candidate templates in a template set corresponding to the current block; according to the first template, determining a block vector candidate list corresponding to the current block; and according to the block vector candidate list corresponding to the current block, determining a block vector corresponding to the current block, and determining a predicted value of the current block according to the block vector.