Motion Vector Predictor Derivation in HEVC Merge Modes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The current HEVC design for motion vector predictor derivation in video compression is computationally complex, particularly in Merge modes, due to the generation and processing of a large number of predictors, which increases encoding and decoding complexity.
Innovation Solution
The method involves processing two distinct subsets of motion information predictors concurrently to derive a final set of predictors, reducing computational complexity by parallel processing and removing duplicates within these subsets, thereby reducing the number of cycles required for motion vector derivation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large number of motion information predictors are generated and processed in HEVC Merge modes, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The set of motion information predictors is divided into two distinct subsets: a first subset containing predictors derived from spatial neighboring blocks, and a second subset containing predictors derived from temporal reference blocks. This segmentation allows the encoder to process and evaluate predictors in organized groups, improving prediction accuracy while managing computational complexity through structured processing.
Solution Approach 2:
Motion information predictors are derived and prepared in advance from neighboring blocks and reference frames before the actual encoding decision is made. The encoder pre-computes multiple candidate predictors and stores them in subsets, allowing the selection process to efficiently choose the best predictor without performing complex computations during the encoding decision phase.
2Measurement precision
If multiple motion information predictors are processed sequentially, then processing accuracy is maintained, but encoding time increases
Solution Approach 1:
The encoder periodically alternates between processing the first subset of spatial predictors and the second subset of temporal predictors in a systematic sequence. This periodic processing pattern ensures that both spatial and temporal prediction information is evaluated thoroughly while maintaining a predictable and efficient processing rhythm that reduces overall encoding time.
3Device complexity
If duplicate motion information predictors are removed from the predictor sets, then computational complexity is reduced, but the number of available predictors decreases
Solution Approach 1:
Duplicate motion information predictors are identified and discarded from the predictor subsets to eliminate redundant computations. The encoder maintains the essential diversity of predictors by keeping one instance of each unique predictor and discarding duplicates, thereby reducing computational complexity while preserving the full range of prediction options needed for accurate encoding.
Data Source
AI summary
A method and device for deriving at least one motion information predictor for encoding of an image portion by motion compensation. At least two distinct first and second subsets of motion information predictors of a first set of motion information predictors are provided. Processing of the first subset of motion information predictors and processing of the second subset of motion information predictors is performed to obtain a final set of motion information predictors usable for predicting said image portion from the reference image portion. Processing of the second subset may comprise removing duplicates from among the motion information predictors of said second subset, and may be performed so as to exclude temporal predictors. At least part of the processing of the second subset of motion information predictors may be performed concurrently with at least part of the processing of the first subset of motion information predictors.