Hierarchical Segmentation for Trajectory Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for estimating trajectories of objects with shape or scale changes, such as a person, face challenges in accurately determining motion vectors due to lack of appropriate movement models, especially in regions with uniform texture or striped patterns.
Innovation Solution
A trajectory estimation method that performs hierarchical segmentation of images to generate subregions across different scales, allowing for the estimation of representative trajectories by searching for similar subregions across hierarchical levels, using shape and pixel value information, and applying dynamic programming for optimal path calculation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional block-based motion vector estimation is used, then the method is simple and computationally efficient, but it cannot accurately estimate trajectories in uniform regions or regions with striped patterns
Solution Approach 1:
The patent applies segmentation by dividing the image into multiple blocks and further segmenting each block into sub-blocks. This hierarchical segmentation allows the system to handle different regions with appropriate granularity, improving trajectory estimation accuracy in uniform regions while maintaining computational efficiency through selective processing.
Solution Approach 2:
The patent implements local quality by using different motion estimation strategies for different regions. Specifically, it identifies uniform regions and striped pattern regions, then applies specialized motion vector estimation methods tailored to each region's characteristics, thereby improving overall accuracy without uniformly increasing complexity across the entire image.
2Measurement precision
If hierarchical segmentation across multiple levels is performed, then trajectory estimation accuracy for objects with shape changes improves, but computational complexity increases
Solution Approach 1:
The patent applies dynamics by making the segmentation hierarchy adaptive rather than fixed. It dynamically adjusts the number of hierarchical levels and the segmentation strategy based on object characteristics, allowing accurate tracking of shape-changing objects while avoiding unnecessary computational overhead in regions where simple motion suffices.
Solution Approach 2:
The patent implements parameter changes by varying segmentation parameters such as block size, number of hierarchical levels, and motion model complexity based on local image characteristics. This allows the system to optimize the balance between accuracy and computational complexity by adjusting parameters according to the specific requirements of different regions and objects.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A trajectory estimation apparatus includes: an image acceptance unit (101) which accepts images that are temporally sequential and included in the video; a hierarchical subregion generating unit (103) which generates subregions at hierarchical levels by performing hierarchical segmentation on each of the images accepted by the image acceptance unit (101) such that, among subregions belonging to hierarchical levels different from each other, a spatially larger subregion includes spatially smaller subregions; and a representative trajectory estimation unit (104) which estimates, as a representative trajectory, a trajectory, in the video, of a subregion included in a certain image, by searching for a subregion that is most similar to the subregion included in the certain image, across hierarchical levels in an image different from the certain image.