Gradient-Parameter Optical Flow for Compression-Efficient Inter Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in achieving high compression ratios with minimal sacrifice in picture quality, particularly in scenarios with limited network bandwidth and memory resources.
Innovation Solution
A bidirectional optical flow prediction method involving obtaining an initial motion vector pair, calculating gradient parameters, and using a piecewise sign function to derive sample and block optical flow parameters for improved inter prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video compression techniques are used, then data transmission efficiency is improved, but picture quality deteriorates
Solution Approach 1:
The patent applies parameter changes by modifying the optical flow calculation process. Instead of using conventional motion estimation, the invention calculates gradient parameters (horizontal and vertical gradients) for each sample position and uses these gradients to determine optical flow parameters. This changes the fundamental parameters of motion estimation, allowing for more accurate motion compensation that preserves picture quality while maintaining compression efficiency.
Solution Approach 2:
The patent substitutes the traditional mechanical motion estimation system with an optical flow-based system. By replacing conventional block-matching algorithms with gradient-based optical flow calculation, the invention achieves more precise motion estimation. The gradient parameters derived from image intensity changes provide a more accurate representation of motion fields, thereby improving picture quality reconstruction during decompression.
2Manufacturing precision
If advanced motion estimation techniques are applied, then picture quality is improved, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the current block into multiple sub-blocks and calculating gradient parameters for each sub-block independently. This segmentation allows the complex optical flow calculation to be performed on smaller, more manageable units. By processing each sub-block separately with its own gradient parameters, the invention achieves high picture quality through detailed motion estimation while reducing overall computational complexity through divide-and-conquer strategy.
3Measurement precision
If gradient parameters are calculated for each sample, then motion estimation accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies preliminary action by calculating gradient parameters for each sample position before performing the optical flow computation. By pre-computing the horizontal and vertical gradients at each sample location, the invention prepares the necessary data in advance. This preliminary calculation of gradient parameters allows the subsequent optical flow derivation to proceed more efficiently, as the gradient values are already available, thereby reducing overall processing time while maintaining high motion estimation accuracy.
Data Source
AI summary
A bidirectional optical flowing prediction method includes obtaining an initial motion vector pair for a current block, obtaining a forward and a backward prediction block according to the forward motion vector and a backward prediction block according to the initial motion vector pair, and calculating gradient parameters for a current sample in the current block. The method further includes obtaining at least two sample optical flow parameters, including a first parameter and a second parameter, for the current sample based on the gradient parameters, obtaining block optical flow parameters based on sample optical flow parameters of samples in the current block, and obtaining a prediction value of the current block. One of the block optical flow parameters is obtained by multiplying the first parameter and a sign function of the second parameter, and the sign function is a piecewise function with at least three subintervals.