Image Processing Circuit Motion Estimation Brightness Shimmer
Solution Overview
Problem
Existing motion estimation and motion compensation technologies, such as those using the sum of absolute differences (SAD) method, are prone to inaccuracies due to brightness changes, leading to shimmer effects in frame rate conversion, especially in areas with dense and irregular textures.
Innovation Solution
An image processing method and circuit that perform low-pass filtering, calculate characteristic values, and blend these with SAD results to estimate motion vectors, reducing the impact of brightness changes and improving vector accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If SAD-based motion estimation is used, then the motion vector estimation is simple to compute, but the accuracy deteriorates due to brightness changes causing shimmer effects
Solution Approach 1:
The patent merges SAD (Sum of Absolute Differences) with characteristic value differences (such as mean brightness or edge strength) to create a hybrid motion estimation metric. This combination allows the system to leverage both the computational efficiency of SAD and the brightness-invariance of characteristic value methods, thereby maintaining accuracy while controlling complexity
Solution Approach 2:
The patent introduces characteristic values (such as mean brightness, variance, or edge density) as additional parameters to the motion estimation process. By computing and comparing these characteristic values across frames, the system can identify and weight regions that are sensitive to brightness changes, thereby improving motion vector accuracy without completely replacing the SAD computation
2Productivity
If conventional motion estimation is used, then processing speed is maintained, but image quality deteriorates due to shimmer artifacts in interpolated frames
Solution Approach 1:
The patent performs preliminary computation of characteristic values (such as mean brightness, variance, or gradient magnitude) for each frame before the actual motion estimation process. These pre-computed values are then used to guide or adjust the SAD-based motion estimation, allowing the system to maintain real-time processing speeds while improving the accuracy of motion vectors in brightness-sensitive regions
Solution Approach 2:
The patent introduces characteristic value-based weighting as an intermediary layer between traditional SAD motion estimation and the final motion vector output. This intermediary mechanism allows the system to selectively adjust motion estimation results in regions where brightness changes are detected, thereby reducing shimmer artifacts without completely redesigning the motion estimation pipeline
Data Source
AI summary
An image processing method and an image processing circuit are provided. The method and circuit are applied to motion estimation. The method includes the steps of: performing low-pass filtering on a first image and a second image, wherein the first image is part of a first frame, the second image is part of a second frame, and the first frame is different from the second frame; calculating a first characteristic value of the first image and calculating a second characteristic value of the second image; calculating a sum of absolute difference (SAD) between the first image and the second image; blending the difference between the first characteristic value and the second characteristic value and the SAD to generate a blended result; and estimating a motion vector between the first image and the second image according to the blended result.