Adaptive Video Stabilization Using Optical Flow and Motion Solve
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Videos captured by image capture devices in motion often appear shaky due to inaccurate determination of device motion, leading to inadequate stabilization.
Innovation Solution
A system that determines optical flow and structure from motion to stabilize videos by selecting between optical flow derived motion and motion solve derived motion, switching techniques based on effectiveness, and combining them for accurate stabilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If optical flow derived motion is used for video stabilization, then the stabilization works well for most scenes, but it fails to provide accurate stabilization in specific conditions such as pure translation or pure rotation scenarios
Solution Approach 1:
The system dynamically switches between optical flow derived motion and motion solve derived motion based on the detected scene conditions. When pure translation or pure rotation is detected, the system transitions from optical flow to motion solve to maintain accurate stabilization, making the stabilization system adaptive to different scene conditions rather than relying on a single method.
Solution Approach 2:
The system changes the motion estimation method parameter based on scene conditions. By detecting specific patterns (pure translation, pure rotation) and switching the underlying motion estimation algorithm, the system adapts its parameters to match the scene requirements, improving reliability across diverse conditions.
2Measurement precision
If motion solve derived motion is used for video stabilization, then it provides accurate stabilization for specific conditions, but it is computationally more intensive and may fail when optical flow is sufficient
Solution Approach 1:
The system applies motion solve, the more computationally intensive method, only partially - specifically when scene conditions indicate pure translation or pure rotation. For other conditions, it uses the lighter optical flow method. This selective application of the more intensive method balances accuracy requirements with energy consumption constraints.
Solution Approach 2:
The system uses feedback from scene analysis to determine when to switch between motion estimation methods. By continuously monitoring scene conditions and adjusting the computational approach accordingly, the system optimizes the balance between measurement precision and energy consumption based on actual runtime conditions.
3Device complexity
If a single motion estimation method is used, then the system is simpler to implement, but it cannot provide accurate stabilization across varying scenes and conditions
Solution Approach 1:
The system segments the video processing into different motion estimation approaches based on scene conditions. By dividing the problem into optical flow for general cases and motion solve for specific cases (pure translation, pure rotation), the system maintains implementation simplicity while improving reliability across varying scenes through conditional segmentation of the processing pipeline.
Data Source
AI summary
An image capture device may capture a video while experiencing motion. Motion of the image capture device during video capture may be determined based on optical flow and structure from motion solve of the video. Optical flow derived motion and/or the motion solve derived motion of the image capture device may be selected to perform stabilization of the video.