Simplified Depth Coding for 3D Video Residuals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently coding residual values for depth intra prediction modes, particularly in multi-view and 3D video coding scenarios.
Innovation Solution
The proposed solution involves techniques for coding residual values using a simplified depth coding (SDC) mode, which adaptively applies context adaptive binary arithmetic coding (CABAC) and skips transform and quantization of DC residual values, enabling broader application across various depth intra prediction modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional residual coding with transform and quantization is applied to depth intra prediction modes, then coding precision is maintained, but device complexity and processing overhead increase
Solution Approach 1:
The patent changes the coding parameters by skipping transform and quantization steps for certain depth intra prediction modes (specifically planar and DC modes), thereby reducing processing complexity while maintaining acceptable coding precision through alternative residual coding approaches
Solution Approach 2:
The patent segments the depth intra prediction modes into different categories: modes that undergo full residual coding (transform and quantization) and modes that use simplified residual coding (skipping transform and quantization), allowing selective application based on mode type to balance precision and complexity
2Productivity
If simplified residual coding is applied to all depth intra prediction modes, then processing speed increases, but coding precision deteriorates
Solution Approach 1:
The patent applies different coding quality levels to different regions of the mode space: simplified residual coding (higher speed) is applied to planar and DC modes where it suffices, while conventional residual coding (higher precision) is applied to other modes that benefit from it, optimizing the trade-off locally for each mode type
3Measurement precision
If transform and quantization are applied to DC residual values, then residual data accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent extracts and removes the transform and quantization steps specifically for DC residual values in certain prediction modes, keeping only the essential residual coding operations to reduce computational complexity while maintaining sufficient accuracy through direct residual signaling
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A video coder may perform a simplified depth coding (SDC) mode, including simplified residual coding, to code a depth block according to any of a variety of, e.g., at least three, depth intra prediction modes. For example, the video coder may perform the SDC mode for coding a depth block according to depth modeling mode (DMM) 3, DMM 4, or a region boundary chain coding mode. In such examples, the video coder may partition the depth block, and code respective DC residual values for each partition. In some examples, the video coder may perform the SDC mode for coding a depth block according to an intra prediction mode, e.g., an HEVC base specification intra prediction mode, such as a DC intra prediction mode or one of the directional intra prediction modes. In such examples, the video coder may code a single DC residual value for the depth block.