Image Encoding Prediction Mode Selection via Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The High Efficiency Video Encoding (HEVC) standard faces low encoding efficiency due to the large number of intra-frame prediction modes, which increases calculation amounts and reduces encoding performance.
Innovation Solution
A method for image encoding that determines a set of rough prediction modes, selects prediction modes based on mode costs and types, and encodes blocks using a target prediction mode from a candidate subset, optimizing the number of modes involved in calculations to improve efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the number of intra-frame prediction modes is increased to improve encoding precision, then the manufacturing precision (encoding accuracy) is improved, but the device complexity (calculation amount) increases
Solution Approach 1:
The patent segments the prediction mode selection process into multiple stages: first determining a rough prediction mode from a limited set of candidate modes, then using this rough mode to guide the selection of final prediction modes. This multi-stage segmentation reduces the calculation burden while maintaining encoding accuracy by avoiding exhaustive evaluation of all possible modes.
Solution Approach 2:
The patent performs preliminary action by determining a rough prediction mode before final mode selection. This preliminary rough mode determination narrows down the search space for subsequent detailed optimization, reducing the overall calculation amount while preserving encoding precision through staged refinement.
2Manufacturing precision
If the number of prediction modes is increased to improve encoding precision, then the manufacturing precision (encoding accuracy) is improved, but the productivity (encoding efficiency) decreases
Solution Approach 1:
The patent divides the mode selection into rough prediction stage and fine optimization stage. By segmenting the process, it achieves high encoding accuracy through multi-stage refinement while improving encoding efficiency by reducing the calculation burden in each individual stage compared to exhaustive search.
Solution Approach 2:
The rough prediction mode determination serves as a preliminary action that guides subsequent optimization. This preliminary step improves encoding efficiency by reducing the search space for final mode selection, while maintaining encoding accuracy through staged refinement of the prediction modes.
3Manufacturing precision
If the number of prediction modes is increased to improve encoding precision, then the manufacturing precision (encoding accuracy) is improved, but the use of energy (computing resource occupation) increases
Solution Approach 1:
The patent segments the computationally intensive mode evaluation into rough prediction and fine optimization phases. This segmentation reduces peak computing resource occupation while maintaining encoding accuracy, as each phase operates on a reduced search space compared to exhaustive evaluation of all modes.
Solution Approach 2:
The rough prediction mode determination performs preliminary action by establishing a guided search direction before detailed optimization. This reduces computing resource occupation during the more intensive fine optimization stage, while preserving encoding accuracy through the staged refinement process.
Data Source
AI summary
A set of rough prediction modes including a MPM subset is determined for a code block during image encoding. A first prediction mode having a mode cost less than a first threshold is selected from the set of rough prediction modes, and a second prediction mode having a mode cost less than a second threshold is selected from the MPM subset. A candidate mode subset is determined based on mode types of prediction modes contained in the set of rough prediction modes and a ranking result of mode costs of the prediction modes, when the mode cost of the first prediction mode is different from the mode cost of the second prediction mode. A target prediction mode is determined for the code block from the candidate mode subset. The code block is encoded with the target prediction mode.