Hash-Based Motion Search for Non-Square Video Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards, such as HEVC and future standards like VVC, face challenges in efficiently processing and compressing digital video due to the growing bandwidth demand, particularly because existing hash-based motion search methods are limited to square blocks, which restricts their performance.
Innovation Solution
The implementation of hash-based motion search methods that can handle non-square blocks, use a fixed subset of samples, perform K-pel integer motion vector precision checks, and apply early termination algorithms to improve prediction accuracy and reduce computational complexity, allowing for more efficient video processing and compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If hash-based motion search is applied to square blocks only, then the method is simple to implement, but it cannot handle non-square blocks which limits its applicability
Solution Approach 1:
The hash-based motion search method is extended to support multiple block shapes (square, rectangular, and non-rectangular) by using a unified hash table structure that can store and retrieve motion information for any block configuration, making the method universally applicable to different block types without requiring separate processing logic for each shape
Solution Approach 2:
Non-square blocks are divided into multiple square sub-blocks for hash-based motion search, allowing the method to process complex block shapes by segmenting them into manageable units that can be individually searched and then combined to determine motion information for the entire block
2Measurement precision
If full rate-distortion cost calculation is performed for each coding mode, then prediction accuracy is maximized, but computational complexity increases significantly
Solution Approach 1:
Hash-based motion search is performed as a preliminary step before rate-distortion cost calculation to identify candidate motion vectors and reference blocks. This preliminary action filters out unlikely candidates, allowing the computationally intensive rate-distortion calculation to be performed only on a reduced set of promising options, thus maintaining accuracy while reducing overall complexity
Solution Approach 2:
The rate-distortion cost calculation is performed selectively for only certain coding modes or candidate vectors identified by the hash-based search, rather than exhaustively for all possible modes. This partial action approach achieves sufficient prediction accuracy by focusing computational resources on the most promising candidates
3Measurement precision
If motion search is performed on the entire current block, then motion estimation accuracy is improved, but computational overhead increases
Solution Approach 1:
A fixed subset of representative samples is extracted from the current block to perform hash-based motion search. These extracted samples capture the essential motion characteristics of the entire block, enabling accurate motion estimation with significantly reduced computational overhead compared to processing all pixels in the block
Data Source
AI summary
Methods, systems and device for hash-based motion estimation in video coding are described. An exemplary method of video processing includes determining, for a conversion between a current block of a video and a bitstream representation of the video, motion information associated with the current block using a hash-based motion search, a size of the current block being M×N, M and N being positive integers and M being not equal to N, applying, based on the motion information and a video picture comprising the current block, a prediction for the current block, and performing, based on the prediction, the conversion.