Motion Vector Prediction Standardization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in reducing the software and hardware implementation difficulty of motion vector prediction, particularly due to the complexity and discrepancies between various prediction modes.
Innovation Solution
A method for motion vector prediction that involves acquiring a candidate list of candidate motion vectors for a current coding block in a preset prediction mode, performing a motion search to obtain an initial optimal motion vector, and then using multiple pixel accuracies to refine this vector and determine a final optimal motion vector, ensuring consistent pixel accuracies and processing loops across prediction modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple prediction modes (conventional AMVP, affine AMVP, intra block copy AMVP, Hash AMVP) are used to improve prediction accuracy, then the prediction performance is enhanced, but the software and hardware implementation difficulty increases due to separate processing loops required for each mode
Solution Approach 1:
The patent creates a unified processing loop that can handle multiple prediction modes (conventional AMVP, affine AMVP, intra block copy AMVP, Hash AMVP) through a single standardized procedure. The candidate list construction and motion search process is made mode-agnostic, allowing the same hardware/software implementation to serve multiple prediction functions without requiring separate processing loops for each mode.
Solution Approach 2:
The patent standardizes the pixel accuracy to 1/16th across all prediction modes and unifies the candidate list construction process. This homogenization eliminates the need for mode-specific adjustments and simplifies the implementation by ensuring that all prediction modes operate under the same parameters and procedures, reducing the complexity of maintaining multiple separate processing loops.
2Adaptability or versatility
If separate processing loops are constructed for each prediction mode to handle their discrepancies, then each mode can be optimized independently, but the overall system complexity and implementation difficulty increases
Solution Approach 1:
The patent designs a universal processing loop that accommodates all prediction modes through standardized candidate list construction and motion search procedures. The system maintains adaptability to different modes while using a single unified loop, eliminating the need for multiple separate loops and reducing the complexity associated with managing mode-specific processing paths.
Solution Approach 2:
The patent merges the previously separate processing loops for different prediction modes into a single unified loop. By combining the candidate list construction, motion search, and compensation processes into one standardized flow that handles all modes, the patent reduces the overall system complexity while preserving the ability to support multiple prediction modes.
Data Source
AI summary
The present disclosure relates to a motion vector prediction method, a video coding method, and related apparatuses and devices. The motion vector prediction method includes: acquiring a candidate list of a current coding block in a preset prediction mode, where the candidate list includes a plurality of candidate motion vectors; performing a motion search of the plurality of candidate motion vectors in the candidate list to obtain an initial optimal motion vector of the current coding block in the preset prediction mode; performing motion compensation on the initial optimal motion vector by using each of a first preset number of pixel accuracies, to determine a final optimal motion vector of the current coding block, where the preset prediction mode comprises any one of a plurality of prediction modes, and pixel accuracies used in any one of the prediction modes are identical.