Motion Vector Refinement for Small-Object Frame Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods result in the disappearance of small objects during frame scaling, affecting image display quality due to reduced frame size, especially in motion estimation and compensation processes.
Innovation Solution
An image processing method and circuit that marks small objects likely to disappear during frame scaling by determining differences in motion vectors and marking them, followed by weighted calculations to generate accurate final motion vectors, ensuring object visibility in interpolated frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the frame size is reduced for motion estimation and compensation, then the hardware cost is reduced and processing efficiency is improved, but small objects in the frame disappear, affecting image display quality
Solution Approach 1:
The patent performs motion estimation on the original high-resolution frame before scaling down, obtaining initial motion vectors. This preliminary action ensures that motion information of small objects is captured before they potentially disappear in the scaled-down version, thereby preserving image display quality while still allowing subsequent scaling for efficient processing
Solution Approach 2:
The patent introduces an intermediary process where motion vectors obtained from the scaled-down frame are adjusted and refined using the preliminary motion vectors from the original frame. This intermediary adjustment step acts as a bridge, combining the processing efficiency of scaled-down frames with the quality preservation of original-frame motion estimation
2Speed
If motion estimation is performed on scaled-down frames, then processing speed is improved, but motion vector accuracy decreases, affecting interpolation quality
Solution Approach 1:
The system performs preliminary motion estimation on the original frame to obtain initial motion vectors before scaling down. This preliminary action provides accurate reference motion information that compensates for the reduced accuracy in scaled-down frame estimation, maintaining overall motion vector precision while enabling faster processing
Solution Approach 2:
The patent implements a feedback mechanism where motion vectors from scaled-down frames are compared and adjusted based on preliminary motion vectors from the original frame. This feedback loop corrects accuracy losses from scaling, ensuring high-quality motion compensation while maintaining processing speed benefits
Data Source
AI summary
The present invention provides an image processing method, which includes the steps of: receiving an image signal, wherein the image signal includes a frame; performing motion estimation on multiple blocks of the frame to generate multiple first motion vectors of the multiple blocks; performing a scaling-down operation on the frame to generate a first scaled-down frame; performing the motion estimation on a first specific block of the first scaled-down frame to generate a second motion vector corresponding to the first specific block; determining differences between the second motion vector of the first specific block and the multiple first motion vectors of the multiple blocks to generate multiple determination results, respectively; and determining multiple final motion vectors of the multiple blocks of the frame according to the multiple determination results, the multiple first motion vectors and the second motion vector.