Depth-Weighted View Synthesis Distortion for 3D Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding methods fail to efficiently evaluate view synthesis distortion in 3D video, particularly in terms of depth impact, leading to suboptimal compression and video quality degradation.
Innovation Solution
A method for estimating view synthesis distortion (VSD) is developed, which involves obtaining a depth map value and determining a weighting factor to calculate VSD as a function of absolute horizontal texture gradients weighted by depth distortion, improving the accuracy of distortion measurement and coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional video coding schemes are used, then compression is achieved, but view synthesis distortion is not properly evaluated, leading to video quality degradation
Solution Approach 1:
The patent introduces a depth-weighted distortion metric that changes the evaluation parameters from conventional 2D distortion measures to 3D depth-aware distortion measures. By incorporating depth map values and computing depth-weighted gradients, the system transforms the distortion evaluation to account for the third dimension, thereby improving measurement precision of view synthesis distortion while preventing video quality loss.
2Productivity
If depth impact is not considered in distortion metrics, then computation is simpler, but 3D video coding efficiency is suboptimal
Solution Approach 1:
The patent performs preliminary computation of depth-weighted gradients and distortion metrics during the encoding process. By pre-computing depth maps and their gradients, and incorporating them into the rate-distortion optimization before actual video coding, the system improves 3D video coding efficiency without excessively increasing device complexity during decoding.
Solution Approach 2:
The patent applies local quality weighting by computing distortion metrics differently for different depth regions. The depth-weighted gradient computation assigns different weights to different spatial locations based on their depth values, allowing the system to focus computational resources on regions where depth distortion has the most impact on perceived quality, thereby improving coding efficiency with controlled complexity.
3Measurement precision
If conventional distortion metrics are used, then computation is faster, but depth distortion impact is not accurately captured
Solution Approach 1:
The patent implements partial computation of depth-weighted distortion metrics by focusing on horizontal gradients and depth-weighted components that have the most significant impact on view synthesis quality. Rather than computing all possible distortion measures, the system selectively computes the most relevant depth-aware metrics, thereby improving measurement precision of depth distortion while limiting the increase in computation time.
Data Source
AI summary
System and method embodiments are provided for achieving improved View Synthesis Distortion (VSD) calculation and more accurate distortion estimation of encoded video frames. An embodiment method includes obtaining a depth map value for a video frame and determining a weighting factor for depth distortion in accordance with the depth map value. The weighting factor maps a pixel range of the depth map value to an output function having higher values for closer image objects and lower values for farther image objects. The VSD for the video frame is then calculated as a function of absolute horizontal texture gradients weighted by a depth distortion value and the weighting factor determined in accordance with the depth map value.