Video Block Decoding with Multiple Template-Derived Intra Modes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video coding methods using template-based intra mode derivation (TIMD) are inefficient as they rely on a single prediction mode, leading to suboptimal coding efficiency for predicting target blocks.
Innovation Solution
A method and device for decoding video data that involves selecting multiple intra candidate modes based on template predictions, generating template predictions for each mode, and reconstructing blocks using a combination of these modes, with cost values and weighting parameters to enhance prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single prediction mode is used in TIMD, then the device complexity is reduced, but the coding efficiency deteriorates
Solution Approach 1:
The patent segments the single prediction mode into multiple candidate modes (typically 2-3 modes). Instead of using one TIMD mode, the encoder generates multiple candidate predictions by dividing the template into different regions and applying different prediction algorithms to each region, then selects the best combination. This segmentation resolves the contradiction by maintaining manageable complexity through structured division while improving coding efficiency through multiple prediction options.
Solution Approach 2:
The patent merges multiple prediction modes into a combined prediction result. After generating multiple candidate predictions from different template regions and modes, the patent combines these predictions (e.g., through weighted averaging or selection) to form the final prediction. This merging approach improves coding efficiency by leveraging the strengths of multiple modes while keeping device complexity controlled through systematic combination rules.
2Manufacturing precision
If multiple intra modes are selected and combined, then the manufacturing precision of prediction is improved, but the device complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-defining a limited set of candidate modes and their combination rules before the actual prediction process. The encoder pre-processes the template to identify potential prediction regions and modes, then uses these pre-established candidates in the main prediction step. This preliminary action improves prediction accuracy by carefully selecting multiple modes while controlling complexity through pre-established selection criteria.
Solution Approach 2:
The patent introduces dynamics by adaptively selecting which candidate modes to use based on the specific characteristics of the current block and template. Rather than always using the same fixed set of modes, the system dynamically adjusts the number and type of candidate modes based on local image features, template quality, and prediction performance. This dynamic adaptation improves prediction accuracy for diverse content while managing complexity through conditional selection.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of decoding a bitstream by an electronic device is provided. The electronic device receives the bitstream. In addition, the electronic device determines a block unit from an image frame according to the bitstream and selects a plurality of intra candidate modes from a plurality of intra default modes for the block unit. The electronic device further generates a template prediction for each of the plurality of intra candidate modes, selects a plurality of prediction modes from the plurality of intra candidate modes based on the template predictions, and reconstructs the block unit based on the plurality of prediction modes.