Occlusion Detection in Video Motion Fields
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion estimation algorithms in video processing struggle to accurately estimate motion in occluded regions and complex movements, leading to errors and artifacts in applications like video compression, 3D rendering, and frame rate up conversion.
Innovation Solution
A processor-implemented method that detects occlusions by comparing motion vectors between frames and uses motion priors to disambiguate between occlusions and errors, allowing for correction of occlusion and error motion vectors, thereby reducing artifacts in video processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If motion estimation algorithms use motion smoothness constraints to solve ambiguous regions, then processing complexity is reduced, but measurement precision of motion vectors deteriorates
Solution Approach 1:
The patent segments the video processing into distinct stages: occlusion detection using multiple motion vector comparisons, occlusion region identification, and separate error handling. This segmentation allows each stage to optimize for its specific function without compromising overall accuracy.
Solution Approach 2:
The patent performs preliminary occlusion detection by comparing motion vectors from multiple frames before final motion estimation. By identifying occluded regions in advance using motion vector consistency checks across multiple frames, the system prepares correction strategies beforehand, improving final motion vector accuracy without excessive processing complexity during main estimation.
2Difficulty of detecting and measuring
If motion estimation algorithms detect occlusions based on motion fields, then occlusion detection capability is improved, but reliability of motion vectors in erroneous regions deteriorates
Solution Approach 1:
The patent implements feedback mechanisms where motion vectors from multiple frames are compared and validated. Erroneous motion vectors are detected through consistency checks, and corrections are applied based on feedback from surrounding blocks and temporal comparisons, improving overall motion vector reliability.
Solution Approach 2:
The patent introduces intermediary steps between motion estimation and final output: occlusion detection acts as an intermediary that identifies problematic regions, and error handling mechanisms serve as intermediaries that correct erroneous motion vectors before final use, thereby maintaining reliability.
3Measurement precision
If algorithms handle complex and fast movement with detailed processing, then motion estimation accuracy is improved, but processing burden increases
Solution Approach 1:
The patent applies partial processing by focusing detailed analysis only on regions identified as occluded or potentially erroneous, rather than processing the entire frame with maximum detail. This selective approach maintains accuracy for critical regions while reducing overall processing burden.
Solution Approach 2:
The patent applies different processing qualities to different regions: occluded regions receive enhanced processing with multiple motion vector comparisons and error handling, while non-occluded regions use standard motion estimation. This local quality differentiation maintains accuracy where needed while reducing overall processing burden.
Data Source
AI summary
Methods, devices and computer-readable mediums for detecting occlusions which occur due to foreground object movement with respect to a background between first and second successive frames. Occlusion detection may use motion estimation with respect to at least a third frame temporally preceding the first frame. Occlusion detection may be based on one or more assumptions such as: occlusion motion vectors are different than other background motion vectors; occlusion motions are likely to be similar to foreground occluding motion; and/or motion estimation will match an occlusion block with a block belonging to a common background object. Occlusion detection may be combined with motion error detection based on a motion field divergence using a motion vector assigned to the occlusion, e.g., for generating an intermediate frame in frame up rate conversion (FRUC).