Fast Fractional Motion Estimation in Video Encoders
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Motion estimation in video encoding systems, particularly in H.264/ISO MPEG-4 AVC encoders, is computationally complex due to the need for multiple references and block sizes, with fractional sample motion estimation becoming a significant bottleneck despite existing fast integer sample motion estimation schemes.
Innovation Solution
A fast fractional sample motion estimation scheme that reduces complexity by examining specific subpixel positions based on the relationship between the best and second-best locations, using diamond patterns and adaptive resolution refinement, and applying early termination criteria to minimize the number of positions to test.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fractional sample motion estimation is performed to improve motion estimation accuracy, then measurement precision is improved, but device complexity increases significantly
Solution Approach 1:
The motion estimation process is divided into two distinct stages: integer sample motion estimation followed by fractional sample refinement. This segmentation allows the system to first perform a coarse search at integer positions, then focus computational resources only on refining the best candidate positions by fractional amounts, thereby reducing overall complexity while maintaining accuracy
Solution Approach 2:
Integer sample motion estimation is performed as a preliminary step before fractional sample refinement. This preliminary action identifies candidate motion vectors at integer positions, which then serve as starting points for the subsequent fractional refinement stage, eliminating the need to search all possible fractional positions from scratch
2Loss of information
If exhaustive fractional sample motion estimation is performed to improve measurement precision, then loss of information is reduced, but productivity decreases
Solution Approach 1:
Fractional sample motion estimation is applied selectively only to the best candidate integer positions identified in the first stage, rather than performing exhaustive search at all positions. This local application of fractional refinement maintains compression efficiency by focusing computational effort where it matters most, while preserving encoding speed
Solution Approach 2:
Instead of performing complete fractional sample motion estimation at all possible positions, the invention applies partial fractional refinement only to a limited set of candidate positions (typically the best integer match and a few neighbors). This partial action is sufficient to achieve near-optimal compression efficiency while dramatically improving encoding speed
Data Source
AI summary
Methods of using motion estimation techniques with video encoders to provide significant data compression with respect to video signals so that the video signals may subsequently be reconstructed with minimal observable information loss. Methods include a fast fractional motion estimation scheme.