Region-of-Interest 3D Video Coding Using Depth Map Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional 3D video compression methods do not efficiently utilize the human vision system's sensitivity to interesting regions of a picture, leading to suboptimal compression performance and increased bandwidth requirements for 3D video data.
Innovation Solution
A region-of-interest based 3D video coding framework that partitions a video picture into regions-of-interest and regions-of-non-interest using depth maps and camera parameters, applying different coding settings to these regions for enhanced compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If conventional 3D video compression methods are used, then compression is applied uniformly across the entire video picture, but this fails to utilize the human vision system's sensitivity to interesting regions, leading to suboptimal compression performance and increased bandwidth requirements
Solution Approach 1:
The video picture is segmented into multiple regions based on depth information and camera parameters. Regions-of-interest (ROIs) are identified and separated from non-interest regions, allowing different compression strategies to be applied to different segments. This segmentation enables the system to allocate more bits to important regions while using more aggressive compression in less important areas, thereby reducing overall bandwidth requirements while maintaining perceived video quality.
Solution Approach 2:
Different compression quality levels are applied to different regions of the video picture based on their importance. Regions-of-interest receive higher quality compression with more detailed encoding, while non-interest regions use lower quality compression. This local quality differentiation exploits the human vision system's sensitivity to certain areas, reducing bandwidth consumption without noticeably degrading the overall viewing experience.
2Productivity
If uniform compression settings are applied to the entire video picture, then the coding process is simple, but this does not account for regional importance and results in inefficient bandwidth utilization
Solution Approach 1:
Before the actual video compression process, preliminary processing is performed to identify and mark regions-of-interest using depth maps and camera parameters. This preliminary action pre-categorizes different regions of the video picture, so that during the subsequent compression phase, the encoder can directly apply appropriate compression settings to each region without complex real-time decisions, thereby improving coding efficiency while managing complexity.
Solution Approach 2:
The compression system dynamically adjusts coding parameters based on the identified regions-of-interest. The compression settings are not static but adapt to the spatial distribution of important regions in each frame. This dynamic approach allows the system to optimize bandwidth utilization for each specific scene composition while maintaining a relatively simple overall coding framework.
Data Source
AI summary
Systems, apparatus, articles, and methods are described including operations for region-of-interest based 3D video coding.