Video Encoder Motion Estimation Using Hierarchical Search Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding methods face challenges in efficiently compressing high-definition/high-resolution video data due to limitations in motion estimation techniques, which affect the quality and efficiency of video compression.
Innovation Solution
A video encoder and its operating method that estimate a motion vector by applying motion estimation criteria, involving the identification of reference blocks, setting search ranges, searching for candidate blocks, and encoding blocks based on estimated motion vectors, while also utilizing a hierarchical structure to improve motion estimation performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If motion estimation techniques are applied to compress video data, then video data size is reduced, but compression efficiency and quality are limited due to existing method constraints
Solution Approach 1:
The patent segments the motion estimation process into multiple stages: initial motion vector candidate generation, refinement search, and final selection. The search space is divided into coarse and fine regions, with different precision levels applied at each stage, enabling efficient compression while maintaining quality.
Solution Approach 2:
The patent introduces hierarchical search dimensions by performing motion estimation at multiple resolution levels. A coarse search is conducted at lower resolution to identify approximate motion vectors, followed by a refined search at full resolution, effectively adding a resolution dimension to the search process.
2Manufacturing precision
If motion estimation is performed to improve video compression quality, then compression quality improves, but computational complexity increases
Solution Approach 1:
The patent performs preliminary motion estimation at a reduced resolution level before the final high-resolution search. This preliminary action provides initial motion vector candidates that guide the subsequent refined search, reducing the overall computational complexity while maintaining compression quality.
Solution Approach 2:
The patent applies partial search strategies by limiting the refined search to a reduced search range around the preliminary motion vector candidates. This partial action concentrates computational resources on the most promising regions, achieving high compression quality without exhaustive full-resolution searching.
3Measurement precision
If search range is expanded to improve motion estimation accuracy, then motion estimation accuracy improves, but processing time increases
Solution Approach 1:
The patent segments the search range into an initial large search range for coarse motion vector estimation, followed by a refined smaller search range around the candidate vectors. This segmentation enables accurate motion estimation while limiting the time-consuming full-range search at high precision.
Solution Approach 2:
The patent adds a resolution dimension to the search process by performing coarse searches at low resolution and refined searches at high resolution. This dimensional approach allows comprehensive search coverage without proportionally increasing processing time, as the majority of the search space is explored at lower computational cost.
Data Source
AI summary
A video encoding method of estimating a motion vector of a current block in a current frame includes identifying a reference block in a reference frame, a first location of the current block within the current frame corresponding to a second location of the reference block within the reference frame, setting a search range in the reference frame, searching for candidate blocks from the search range of the reference frame, each of the candidate blocks satisfying a corresponding first motion estimation criterion of first motion estimation criteria, searching for a prediction block from among the candidate blocks, the prediction block satisfying a second motion estimation criterion, estimating, as the motion vector of the current block, a motion vector corresponding to a displacement between the reference block and the prediction block, and encoding the current block based on the estimated motion vector.