Motion Estimation Using Momentum and Acceleration Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional block-based motion estimation systems often rely on randomness to generate candidate motion vectors, which can lead to slow convergence and incorrect motion representation, especially in rapidly changing scenes, as they do not effectively consider the physical nature of real-world objects.
Innovation Solution
The method involves calculating candidate motion vectors based on the physical momentum and acceleration of objects, predicting their future positions and velocities, and storing these vectors at offset locations to improve tracking and adaptation to changing motion, thereby reducing the reliance on random candidates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If random candidate motion vectors are used in conventional block-based motion estimation, then the system can cover a wide range of possible motions, but the convergence speed is slow and the accuracy of true motion representation is reduced
Solution Approach 1:
The system performs preliminary action by calculating momentum and acceleration vectors from historical motion data before the motion estimation search. These pre-calculated physical vectors are used to generate initial candidate motion vectors, providing a head start that guides the search towards the true motion direction and significantly reduces convergence time while improving accuracy.
Solution Approach 2:
The system changes the parameters used for candidate generation from purely random or spatially-correlated vectors to physically-based vectors incorporating momentum and acceleration. This parameter change transforms the candidate distribution to be more aligned with actual object motion dynamics, improving both convergence speed and motion estimation accuracy.
2Adaptability or versatility
If random perturbation is used to adapt to changing motion, then the system can eventually converge to new motion patterns, but it may become stuck in local minima or converge too slowly
Solution Approach 1:
The system applies dynamics by incorporating momentum and acceleration calculations that naturally adapt to changing motion patterns. Instead of static random perturbation, the dynamic physical vectors continuously adjust to reflect current object motion behavior, enabling the system to quickly adapt to motion changes while maintaining high convergence speed through physics-based guidance.
Solution Approach 2:
The system uses feedback from historical motion vectors to calculate momentum and acceleration, which then feed back into the candidate generation process. This feedback mechanism ensures that the candidate vectors continuously adapt to the latest motion patterns, preventing the system from getting stuck in local minima and enabling rapid convergence to the true motion.
3Productivity
If a small number of candidate motion vectors is used to minimize computational expense, then the processing speed is improved, but the ability to identify true motion is reduced
Solution Approach 1:
The system changes the quality parameters of candidate vectors from random to physics-based, incorporating momentum and acceleration. This parameter change increases the information content and relevance of each candidate vector, allowing the system to achieve high motion identification accuracy with fewer candidates, thus maintaining processing speed while improving precision.
Solution Approach 2:
The system replaces the mechanical brute-force approach of testing many random candidates with a smarter approach based on physical laws (momentum and acceleration). This substitution reduces the number of candidates needed while maintaining or improving accuracy, as the physics-based vectors are inherently more likely to represent true motion.
Data Source
AI summary
There is provided a method and apparatus for motion estimation in a sequence of video images. The method comprises a) subdividing each field or frame of a sequence of video images into a plurality of blocks, b) assigning to each block in each video field or frame a respective set of candidate motion vectors, c) determining for each block in a current video field or frame, which of its respective candidate motion vectors produces a best match to a block in a previous video field or frame, d) forming a motion vector field for the current video field or frame using the thus determined best match vectors for each block, and e) forming a further motion vector field by storing a candidate motion vector derived from the best match vector at a block location offset by a distance derived from the candidate motion vector. Finally, steps a) to e) are repeated for a video field or frame following the current video field or frame. The set of candidate motion vectors assigned at step b) to a block in the following video field or frame includes the candidates stored at that block location at step e) during the current video field or frame The method enables a block or tile based motion estimator to improve its accuracy by introducing true motion vector candidates derived from the physical behaviour of real world objects.