Global Motion Estimation Using B-Frame Recalculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for global motion estimation in video sequences, particularly using MPEG motion vectors, are hindered by noise in motion vector information, leading to inaccurate motion estimation and poor quality panoramic image generation.
Innovation Solution
The method involves detecting motion estimation failures and recalculating motion estimates along alternative routes using both forward and backward motion vectors in MPEG-encoded frames, allowing for accurate global motion estimation by accumulating individual motion estimations between frames, and selecting the most reliable motion estimation based on predetermined criteria.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional motion vector-based global motion estimation is used, then computation is simplified and storage is reduced, but measurement precision deteriorates due to noise in motion vector information
Solution Approach 1:
The patent segments the global motion estimation problem into multiple local motion estimations between adjacent frames. Instead of directly estimating global motion from noisy motion vectors, the method divides the temporal sequence into frame-by-frame segments, performing local motion compensation and accumulation to achieve more accurate global motion estimation while managing computational complexity through incremental processing
Solution Approach 2:
The patent introduces intermediate reference frames as mediators in the motion estimation process. By using stable reference frames and performing motion estimation through intermediate steps rather than directly between distant frames, the method reduces the impact of noise accumulation and improves measurement precision without requiring excessive computational resources
2Reliability
If motion-compensated inter-frame encoding is used, then storage and transmission efficiency are improved, but reliability deteriorates due to noise in motion vector information affecting panoramic image quality
Solution Approach 1:
The patent performs preliminary motion estimation and reference frame selection before panoramic image generation. By pre-identifying stable reference frames and computing preliminary motion compensations, the method ensures higher reliability in the final panoramic image while minimizing information loss through careful selection and preprocessing of motion data
Solution Approach 2:
The patent implements feedback mechanisms where motion estimation results are evaluated and used to adjust reference frame selection and motion parameter refinement. This iterative feedback process improves reliability by continuously optimizing motion information accuracy based on estimated quality metrics and error detection
Data Source
AI summary
The invention relates to estimating the global motion between frames of a motion-compensated inter-frame encoded video sequence, directly from the motion vectors encoded within the frames. For any particular frame, a motion estimation is determined from motion vectors direct from the frame's anchor frame to the frame in question. This motion estimation is then checked against pre-determined criteria, and where the criteria are not met, re-estimation along a different route is performed, using the bi-directional motion vectors contained within B-frames. A panoramic image generating method and system which makes uses of the global motion estimations thus obtained is also described.