Template Matching for Intra Prediction Mode Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards, such as JVET, face challenges in efficiently decoding video frames without explicit intra prediction mode signaling, leading to increased bitstream overhead and reduced decoding efficiency.
Innovation Solution
A method is introduced where a decoder defines a coding unit (CU) template within a decoded area, generates candidate prediction templates using different intra prediction modes, calculates matching costs, and selects the mode with the lowest cost to generate a prediction CU, thereby omitting explicit intra prediction mode signaling in the bitstream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If explicit intra prediction mode signaling is transmitted in the bitstream, then decoding accuracy is improved, but bitstream overhead increases
Solution Approach 1:
The patent extracts the intra prediction mode information from the bitstream by deriving it implicitly from the template matching process. Instead of transmitting explicit mode signals, the decoder derives the mode by comparing candidate prediction templates with the actual block, effectively taking out the need for explicit signaling while maintaining decoding accuracy.
Solution Approach 2:
The system performs self-service by having the decoder autonomously determine intra prediction modes through template matching without relying on external signaling. The decoder generates candidate templates, compares them with the current block, and selects the best matching mode independently, eliminating the need for bitstream transmission of mode information.
2Measurement precision
If template matching with multiple candidate prediction templates is performed, then intra prediction accuracy is improved, but decoding complexity increases
Solution Approach 1:
The patent segments the decoding process into distinct stages: generating candidate prediction templates using different intra prediction modes, comparing each candidate with the current block, calculating matching costs, and selecting the best mode. This segmentation allows systematic evaluation of multiple candidates while managing complexity through structured processing steps.
Solution Approach 2:
The system performs partial action by evaluating a limited set of candidate prediction templates corresponding to specific intra prediction modes rather than exhaustively searching all possible modes. This selective approach achieves sufficient prediction accuracy while controlling decoding complexity by focusing on the most promising candidates.
3Quantity of substance
If intra prediction mode is omitted from bitstream, then bitstream overhead is reduced, but decoding reliability may worsen
Solution Approach 1:
The patent implements feedback by using the template matching results to verify and validate the derived intra prediction mode. The decoder compares candidate templates with the actual block, calculates matching costs as feedback, and uses this feedback to confirm the reliability of the selected mode, ensuring accurate reconstruction without explicit signaling.
Solution Approach 2:
The system performs preliminary action by pre-generating candidate prediction templates using various intra prediction modes before the actual decoding decision. This preliminary generation of candidates with different modes allows the decoder to evaluate multiple possibilities and select the most reliable one, ensuring decoding reliability is maintained even without explicit mode signaling in the bitstream.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A method of decoding JVET video, comprising defining a coding unit (CU) template within a decoded area of a video frame, the CU template being positioned above and/or to the left of a current decoding position for which data was intra predicted, defining a search window within the decoded area, the search window being adjacent to the CU template, generating a plurality of candidate prediction templates based on pixel values in the search window, each of the plurality of candidate prediction templates being generated using different intra prediction modes, calculating a matching cost between the CU template and each of the plurality of candidate prediction templates, selecting an intra prediction mode that generated the candidate prediction template that had the lowest matching cost relative to the CU template, and generating a prediction CU for the current decoding position based on the intra prediction mode.