Multi-Camera Subject Tracking via Feature Information Exchange
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to effectively track a specific subject using multiple image capture apparatuses with different image capture positions or directions, due to issues with template matching and changes in the subject's image capture region.
Innovation Solution
A system comprising a first and second image capture apparatus, each with different image capture directions, and corresponding control apparatuses that decide and track a specific subject based on feature information from both apparatuses, ensuring accurate tracking even with varying image capture positions or directions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If template matching is used to track a specific subject, then tracking can be performed with simple image capture apparatuses, but the apparatuses must be arranged close together with similar image capture positions or directions
Solution Approach 1:
The patent changes the parameter of subject representation from template-based to feature-based. Instead of using templates that require spatial proximity for matching, the system extracts features (such as facial features, body posture, or distinctive characteristics) from the subject and uses these features for recognition. This allows the image capture apparatuses to be arranged at different positions and directions while still maintaining accurate subject tracking through feature similarity comparison.
2Adaptability or versatility
If the image capture region changes between multiple apparatuses, then more flexible apparatus arrangement is possible, but the similarity of the tracking target subject lowers making it difficult to recognize the same subject
Solution Approach 1:
The patent segments the subject into distinct features rather than treating the entire subject as a single template. By dividing the subject into key features (such as face, body, or distinctive characteristics), the system can recognize the subject even when only parts are visible in different image capture regions. This segmentation approach maintains recognition accuracy regardless of how much of the subject is captured by each apparatus.
Solution Approach 2:
The patent changes from template-based matching to feature-based matching. Instead of requiring the entire subject to be captured and compared as a whole template, the system extracts and compares specific features from the subject. This parameter change allows accurate recognition even when the image capture region varies between apparatuses, as long as the extracted features remain consistent.
3Adaptability or versatility
If the size of the tracking target subject in an image changes between apparatuses, then different capture scenarios are possible, but the similarity of the tracking target subject lowers making it difficult to recognize the same subject
Solution Approach 1:
The patent changes the matching parameter from overall template similarity to feature similarity. Instead of comparing the entire subject image which is affected by size changes, the system extracts and compares specific features from the subject. This feature-based approach is invariant to size changes, allowing consistent subject identification across different capture scenarios where the subject size varies in the captured images.
Data Source
AI summary
A system which includes a first and second image capture apparatuses, and a first and second control apparatuses which control the second image capture apparatus to track a predetermined subject based on one of a first image of the first image capture apparatus and a second image of the second image capture apparatus. The first control apparatus decides the predetermined subject from subjects included in the first image, generates first feature information of a first region of the predetermined subject, and controls the second image capture apparatus to track the predetermined subject. The second control apparatus generates second feature information of a subject included in the second image, decides the predetermined subject based on the first feature information and the second feature information obtained from the first control apparatus, and controls the second image capture apparatus to track the predetermined subject.