Video Coding Regions of Interest Signaling and Gradual Decoder Refresh
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards, particularly HEVC, lack mechanisms for efficient signaling of regions of interest (ROIs) and gradual decoding refresh (GDR), which limits random access and error resilience in video decoding processes.
Innovation Solution
The implementation of techniques that allow for signaling information related to ROIs and GDR through sequence parameter sets, video parameter sets, and supplemental enhancement information (SEI) messages, enabling decoders to identify and process ROI data independently and discard non-relevant slices, thereby enhancing decoding efficiency and error resilience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current video coding standards (HEVC) are used without ROI and GDR mechanisms, then decoding processes are simpler, but random access and error resilience are limited
Solution Approach 1:
The video picture is divided into multiple slices, and each slice is further divided into regions of interest (ROIs) and non-ROI regions. This segmentation allows the decoder to selectively process only relevant regions, improving error resilience and random access capability without requiring complete decoding of the entire picture, thus balancing reliability improvement with manageable complexity increase.
Solution Approach 2:
The patent extracts and separates the signaling information for ROIs and GDR into dedicated data structures and messages (such as SEI messages). This extraction allows the decoder to identify and process only the necessary regions while discarding non-relevant data, thereby improving error resilience and random access without proportionally increasing overall decoding complexity.
2Productivity
If all slices are decoded completely, then video quality is maintained, but decoding efficiency decreases when only certain regions are needed
Solution Approach 1:
The patent applies different decoding strategies to different regions of the picture. ROIs are decoded with high quality using complete slicing and progressive refinement, while non-ROI regions can be decoded with lower quality or discarded entirely. This local quality differentiation improves overall decoding efficiency when only certain regions are needed, while maintaining video quality in the important regions.
Solution Approach 2:
The decoder performs partial decoding by selectively processing only the slices and regions that are necessary for the current access requirement. When random access is needed, the decoder can stop after decoding only up to the required region rather than completing the entire picture, thus improving productivity without excessive loss of information in the decoded regions.
3Reliability
If signaling information for ROIs and GDR is added to the bitstream, then random access and error resilience improve, but bitstream complexity increases
Solution Approach 1:
The patent introduces SEI (Supplemental Enhancement Information) messages as an intermediary layer to carry ROI and GDR signaling information. These SEI messages are separate from the core video data structure, allowing the addition of random access and error resilience capabilities without fundamentally complicating the main bitstream structure. The intermediary SEI layer provides a clean interface for adding functionality while maintaining compatibility with existing decoding frameworks.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
During a coding process, systems, methods, and apparatus may code information indicating whether gradual decoder refresh (GDR) of a picture is enabled. When GDR is enabled, the coding process, systems, methods, and apparatus may code information that indicates whether one or more slices of the picture belong to a foreground region of the picture. In another example, during a coding process, systems, methods, and apparatus may decode video data corresponding to an ISP identification (ISP ID) for one of the ISPs for slices of a picture. The systems, methods, and apparatus may decode video data corresponding to an ROI using the ISP.