Video Anti-Shaking Using Global Relative Motion Vector Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video anti-shaking techniques struggle to distinguish between camera shaking and intended motion, leading to suboptimal anti-shaking effects in video recordings.
Innovation Solution
A method and device that calculate and filter global relative motion vectors to differentiate between shaking and intended motion, using a combination of lens motion detection and motion estimation algorithms to compensate for camera shake by cropping the video frame effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If optical anti-shaking techniques are used to detect and compensate motion, then shaking is reduced, but the system complexity increases due to gyro sensors and lens adjustment mechanisms
Solution Approach 1:
The patent replaces the mechanical/optical anti-shaking system (gyro sensors, lens adjustments) with a digital signal processing approach. Motion vectors are calculated from video frames and used to stabilize the video through digital cropping and composition, eliminating the need for complex mechanical stabilization components.
2Stability of the object's composition
If electronic anti-shaking techniques process captured videos to reduce shaking, then shaking is reduced, but the processing time and computational resources increase
Solution Approach 1:
The patent calculates motion vectors between consecutive video frames during the video capture phase rather than after complete recording. This preliminary calculation of global relative motion vectors enables real-time or near-real-time stabilization processing, reducing the overall processing time required.
3Measurement precision
If motion estimation algorithms are used to calculate global relative motion vectors, then shaking compensation accuracy is improved, but the computational complexity increases
Solution Approach 1:
The patent segments the video into discrete frames and calculates motion vectors between adjacent frames independently. This frame-by-frame approach breaks down the complex task of analyzing entire video sequences into manageable units, reducing computational complexity while maintaining accuracy through sequential processing.
Data Source
AI summary
A video anti-shaking method and a video anti-shaking device are provided. The method includes: obtaining a first GRMV (global relative motion vector) of a current video frame relative to its previous video frame, and a sequence of second GRMVs of reference video frames relative to their respective previous video frames; calculating a sequence of third GRMVs of the reference video frames relative to the current video frame, respectively, based on the first GRMV and the sequence of second GRMVs; and filtering the sequence of third GRMVs to obtain a motion vector for compensating shaking of the current video frame. Camera shaking and intended motion can be distinguished.