Image Processing Apparatus Periodic Pattern Motion Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies face challenges in accurately detecting repetitive patterns in frames during motion estimation, leading to misjudgments and unnatural fragmented patterns, especially when dealing with periodic patterns of various sizes and ranges.
Innovation Solution
The proposed image processing apparatus and method involve downsizing frames, marking blocks with periodic patterns, performing motion estimations based on these labels, and adjusting motion vectors to correct for repetitive patterns, thereby improving the accuracy of motion compensation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If motion estimation is performed on frames with repetitive patterns, then motion vectors can be generated for frame rate conversion, but misjudgments occur leading to periodic broken patterns and degraded viewing quality
Solution Approach 1:
The patent applies preliminary action by detecting periodic patterns in the current frame and reference frame before performing motion estimation. The system identifies blocks with periodic features and marks them with labels, then uses these labels to guide the motion estimation process, preventing misjudgments from occurring in the first place
Solution Approach 2:
The patent applies local quality by treating different blocks differently based on their periodic pattern characteristics. Blocks identified as having periodic patterns are marked with special labels and processed with adjusted motion estimation parameters, while non-periodic blocks are processed normally, allowing localized adaptation to different content types
2Measurement precision
If periodic patterns of various sizes and ranges are detected to avoid misjudgments, then motion estimation accuracy improves, but computational burden increases
Solution Approach 1:
The patent applies segmentation by dividing the frame into multiple blocks and further into sub-blocks for periodic pattern detection. This hierarchical segmentation allows the system to detect patterns at different scales (first blocks for larger patterns, n-th blocks for smaller patterns) without having to analyze the entire frame at once, reducing overall computational complexity
Solution Approach 2:
The patent applies partial action by performing motion estimation on marked blocks with periodic patterns using the generated motion vectors as references, rather than performing exhaustive motion estimation on all possible block sizes and ranges. This selective approach reduces computational burden while maintaining accuracy for critical periodic pattern regions
Data Source
AI summary
An image processing apparatus and method are provided. An image processing device is configured to execute the following operations. The apparatus marks a first periodic block in a down-sized current frame with a first label. The apparatus performs a first motion estimation on the down-sized current frame and a down-sized reference frame based on the first label to generate first motion vectors. The apparatus marks an n-th periodic block having another periodic feature in a current frame with an n-th label. The apparatus performs an n-th motion estimation on the current frame and a reference frame based on the n-th label to generate n-th motion vectors. The apparatus performs a motion compensation on the current frame and the reference frame based on the n-th motion vectors to generate a compensated frame.