Video Decoder Motion Vector Estimation for Error Concealment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video signal decoding technologies fail to effectively conceal degradation caused by transmission errors, particularly in inter-frame motion-compensated predictive coding, as concealment vectors are lost when significant portions or entire frames are corrupted, leading to disruption and loss of synchronization.
Innovation Solution
An apparatus and method for decoding video signals using bidirectional motion-compensated prediction, where motion vector estimation generates a substitute motion vector by forming the difference between motion vectors of bidirectionally coded frames, allowing for substitution of frame parts from other decoded frames based on similarity criteria, ensuring continued decoding even with errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If inter-frame motion-compensated predictive coding is used to reduce data rate, then transmission efficiency is improved, but vulnerability to transmission errors increases
Solution Approach 1:
The patent stores multiple anchor frames (I-frames and P-frames) in advance in the frame buffer before decoding occurs. When transmission errors corrupt a frame, the decoder can retrieve previously stored anchor frames to perform error concealment, thus preparing compensatory resources beforehand to mitigate the vulnerability introduced by inter-frame predictive coding.
Solution Approach 2:
The patent introduces motion vector estimation as an intermediary mechanism that operates independently of the transmitted data. By estimating motion vectors from available anchor frames and using them to conceal errors, the system creates a fallback pathway that bypasses corrupted transmitted data, thus reducing error vulnerability while maintaining the efficiency benefits of predictive coding.
2Reliability
If concealment vectors are used to repair corrupted frames, then error concealment is improved, but synchronization loss occurs when significant portions of frames are corrupted
Solution Approach 1:
The patent copies pixel values from previously decoded anchor frames to reconstruct corrupted regions. By copying entire macroblocks or regions from reliable anchor frames rather than relying solely on motion-compensated interpolation, the system maintains synchronization because the copied data comes from already-synchronized reference frames, avoiding the propagation of synchronization errors.
Solution Approach 2:
The patent replaces the mechanical dependency on transmitted concealment vectors with a software-based motion vector estimation system. This substitution allows the decoder to independently generate motion vectors from available anchor frames without relying on potentially corrupted transmitted vectors, thus maintaining synchronization even when significant portions of frames are corrupted.
3Measurement precision
If bidirectional prediction is used for B-frames to improve quality, then prediction accuracy is improved, but complexity of decoding increases
Solution Approach 1:
The patent applies bidirectional prediction selectively rather than uniformly across all B-frames. By determining prediction mode on a per-macroblock basis and using anchor frames as references, the system achieves high prediction accuracy where needed while avoiding the full complexity of bidirectional prediction in all cases, thus balancing quality and complexity.
Solution Approach 2:
The patent makes anchor frames serve multiple functions: they act as reference frames for bidirectional prediction of B-frames, as sources for motion vector estimation during error concealment, and as fallback references when transmission errors occur. This multi-functionality reduces the need for separate error concealment mechanisms, thereby reducing overall decoding complexity while maintaining prediction accuracy.
Data Source
AI summary
A decoder for video signals, such as MPEG which use motion-compensated bidirectional predictive coding, performs concealment of lost or corrupted portions of a picture. For this purpose, it estimates missing motion vectors by combining the two vectors which accompany a bidirectionally coded frame to create a substitute vector. An encoder can be modified to enhance this decoder operation, including forcing at least one frame per group of frames to be coded using bidirectional prediction, and constraining the two vectors so that the substitute vector is closer to the wanted value.