Multi-view Video Codec Residual Prediction Robustness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The current multi-view video codec design for residual prediction in 3D-HEVC is vulnerable to transmission losses, as it relies on the presence of a reference picture (VA) in the decoded picture buffer, making it difficult to detect losses and leading to incorrect classification of coding units for residual prediction.
Innovation Solution
The solution involves checking the availability of reference pictures through parameter sets in the data stream, ensuring that the decision to enable or disable residual prediction is based on the presence of required pictures in the decoded picture buffer, and using flags to indicate the availability of third reference pictures, thereby improving robustness and reducing processing complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the codec checks the presence of reference picture VA in the decoded picture buffer to enable residual prediction, then residual prediction can be performed when available, but transmission losses cannot be detected and lead to incorrect classification of coding units
Solution Approach 1:
The patent introduces an intermediary mechanism by checking the presence of reference picture VA through parameter sets in the data stream rather than directly checking the decoded picture buffer. This intermediary check allows the system to detect transmission losses of reference picture VA while maintaining the residual prediction functionality, thus resolving the contradiction between reliability and loss detection capability
Solution Approach 2:
The patent performs preliminary checks on the availability of reference picture VA through parameter sets before attempting to use it for residual prediction. By checking the presence of required pictures in advance through data stream parameter sets, the system can proactively identify potential transmission losses and adjust the decoding process accordingly, preventing incorrect classification of coding units
2Measurement precision
If the codec checks the decoded picture buffer for reference picture availability, then accurate detection is possible, but processing complexity increases
Solution Approach 1:
The patent extracts the reference picture availability information from the decoded picture buffer and obtains it directly from parameter sets in the data stream. This extraction approach allows the system to maintain accurate detection of reference picture availability while avoiding the complexity of continuously monitoring and checking the decoded picture buffer state
Solution Approach 2:
The patent uses parameter sets in the data stream as a copy or representation of the reference picture availability status. Instead of directly checking the decoded picture buffer, the system reads the availability information from the parameter sets, which contain the necessary information about reference picture presence, thereby reducing processing complexity while maintaining detection accuracy
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Multi-view video coding/decoding supporting residual prediction involving a reference picture of different view and timestamp compared to a currently coded/decoded picture is improved in terms of robustness and/or processing cost by avoiding checking any picture availability or non-availability in a decoded picture buffer.