Video Encoding Prediction Unit Size-Based Motion Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding methods experience increased memory access and bandwidth requirements due to the need for interpolating subpixels during motion compensation, particularly when performing bi-directional motion estimation, which can lead to inefficient data retrieval from reference frames.
Innovation Solution
The method limits memory access by determining whether unidirectional or bi-directional motion estimation modes are used based on the size of the prediction unit, optimizing the interpolation process to reduce the number of pixels read from the reference frame, and employing a hierarchical coding structure to adaptively adjust encoding units for efficient data retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If bi-directional motion estimation is used to improve prediction accuracy, then the prediction precision is improved, but the memory access bandwidth increases significantly
Solution Approach 1:
The prediction unit is divided into multiple sub-blocks, and motion estimation is performed selectively on each sub-block. For small-sized prediction units, only unidirectional motion estimation is performed, while for large-sized prediction units, bi-directional motion estimation is performed. This segmentation allows the system to balance prediction accuracy with memory access efficiency.
Solution Approach 2:
Instead of performing full bi-directional motion estimation on all prediction units, the invention applies partial action by performing bi-directional motion estimation only on large-sized prediction units where it provides significant benefit, while using simpler unidirectional estimation for small-sized units where the overhead would outweigh the benefits.
2Measurement precision
If subpixel interpolation is performed to improve motion compensation accuracy, then the motion compensation precision is improved, but the number of memory accesses increases
Solution Approach 1:
Different interpolation strategies are applied locally based on prediction unit size. For small-sized prediction units, unidirectional interpolation is used which requires fewer memory accesses. For large-sized prediction units, full bi-directional subpixel interpolation is performed to maintain high precision where the additional memory access is more justifiable.
3Measurement precision
If prediction units are made smaller to improve prediction accuracy, then the prediction precision is improved, but the number of motion vectors increases leading to increased processing complexity
Solution Approach 1:
The invention dynamically adjusts the motion estimation strategy based on prediction unit size. Small prediction units use unidirectional motion estimation with simpler processing, while large prediction units use bi-directional motion estimation. This dynamic adaptation allows the system to maintain high prediction accuracy where needed while avoiding unnecessary processing complexity elsewhere.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method and apparatus for encoding and decoding a video. The method of encoding the video includes: determining whether a unidirectional motion estimation mode and a bidirectional motion estimation mode are used based on a size of a current prediction unit to be encoded, performing motion estimation and motion compensation on the current prediction unit by using the determined motion estimation mode, determining a motion estimation mode of the current prediction unit based on an encoding cost of the current prediction unit obtained through the motion estimation and the motion compensation, and encoding determined motion estimation mode information based on the size of the current prediction unit.