Video Coding Using Visual Quality Metrics for Motion Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video coding standards rely on pixel-based distortion metrics, which do not accurately reflect the visual quality of reconstructed frames, leading to suboptimal compression efficiency and visual quality.
Innovation Solution
Incorporating a visual quality evaluation module within the coding loop to assess visual quality using metrics like noise, sharpness, and edge detection, and using this evaluation to guide motion estimation and target coding parameter decisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If pixel-based distortion metrics (SAD, SATD, SSD) are used for motion estimation and mode selection, then coding complexity is reduced and processing is simplified, but visual quality of reconstructed frames deteriorates because pixel-based distortion does not correlate well with actual perceived quality
Solution Approach 1:
The patent changes the parameter used for motion estimation and mode selection from pixel-based distortion metrics (SAD, SATD, SSD) to visual quality metrics that better correlate with human perception. This includes using metrics that account for human visual system characteristics such as sensitivity to different spatial frequencies and contrast masking effects, thereby improving visual quality assessment accuracy while maintaining coding efficiency
Solution Approach 2:
The patent introduces visual quality metrics as an intermediary between pixel-based distortion and perceived quality. These metrics serve as a bridge that translates pixel-level differences into perceptually relevant quality assessments, allowing the coding system to make decisions based on perceived quality rather than raw pixel differences
2Productivity
If conventional pixel-based distortion metrics are used to select video coding modes, then encoding speed is maintained and processing time is reduced, but visual quality of decoded frames deteriorates because minimum pixel distortion does not guarantee best visual quality
Solution Approach 1:
The patent changes the optimization parameter from pixel-based distortion minimization to visual quality maximization. By using visual quality metrics that incorporate human visual system characteristics during mode selection, the system achieves better visual quality without significantly impacting encoding speed, as the metrics are designed to be computationally efficient
Solution Approach 2:
The patent dynamically adjusts coding decisions based on visual quality metrics that adapt to local image characteristics. The system evaluates visual quality in different regions and coding scenarios, dynamically selecting modes that optimize perceived quality rather than using fixed pixel-based thresholds
3Measurement precision
If visual quality evaluation module is added to the coding loop to assess visual quality using human visual system metrics, then visual quality of reconstructed frames is improved, but device complexity and computational load increase
Solution Approach 1:
The patent applies visual quality evaluation selectively to different regions and coding decisions within the coding loop, rather than uniformly to all blocks. This local application reduces overall computational complexity while maintaining visual quality improvements in critical regions where human perception is most sensitive
Solution Approach 2:
The patent implements partial visual quality evaluation, applying it to key decision points such as motion estimation and mode selection, rather than to every coding parameter. This partial application provides sufficient visual quality improvement while avoiding the excessive complexity of comprehensive visual quality optimization at all levels
Data Source
AI summary
One video coding method includes at least the following steps: utilizing a visual quality evaluation module for evaluating visual quality based on data involved in a coding loop; and referring to at least the evaluated visual quality for performing motion estimation. Another video coding method includes at least the following steps: utilizing a visual quality evaluation module for evaluating visual quality based on data involved in a coding loop; and referring to at least the evaluated visual quality for deciding a target coding parameter associated with motion estimation.