Visual Signature Tracking for Appearance Changes Across Cameras
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing tracking devices struggle to accurately track individuals when their appearance changes, leading to false negatives, particularly due to variations in viewpoint and posture across multiple cameras.
Innovation Solution
A method and device that updates a list of reference visual signatures by adding new signatures when a user validates a changed appearance, using thresholds and user input to confirm identity, and employs neural networks for signature determination and comparison.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single reference visual signature is used for tracking, then the device complexity is low, but the reliability of tracking deteriorates when appearance changes occur
Solution Approach 1:
The reference visual signature list is made dynamic by allowing automatic addition of new signatures when appearance changes are detected. The system transitions from a static single-signature model to a dynamic multi-signature model that adapts to appearance variations, resolving the contradiction between reliability and complexity.
Solution Approach 2:
The system changes the parameter of reference signature quantity from one to multiple. By storing and comparing against multiple reference visual signatures instead of a single signature, the system maintains high reliability during appearance changes while managing complexity through efficient distance comparison algorithms.
2Reliability
If multiple reference visual signatures are stored, then the reliability of tracking during appearance changes improves, but the measurement precision of distance comparison deteriorates
Solution Approach 1:
The system applies local quality by using two different threshold values (first threshold and second threshold) for different comparison scenarios. The first threshold is used for initial identification while the second threshold provides a margin for appearance variations, allowing precise local adaptation to different tracking conditions.
3Ease of operation
If automatic addition of visual signatures is implemented, then the ease of operation improves, but the reliability of identification deteriorates due to potential false positives
Solution Approach 1:
The system implements feedback through a dual-threshold mechanism that provides confirmation at different levels. The first threshold provides initial identification feedback, while the second threshold provides a confirmation layer for appearance changes, ensuring reliable identification while maintaining ease of operation through automatic processes.
4Adaptability or versatility
If two threshold values are used for comparison, then the adaptability to appearance changes improves, but the device complexity increases
Solution Approach 1:
The comparison process is segmented into two distinct threshold checks: a first threshold for initial identification and a second threshold for appearance change detection. This segmentation allows the system to handle different scenarios with appropriate thresholds, improving adaptability while keeping processing logic organized and manageable.
Data Source
Figure 1
Figure 2
AI summary
The present invention relates to a method for tracking a reference element (Eref) on images of an environment, the reference element (Eref) being associated with a list of reference visual signatures, the method comprising the steps of: - receiving an image of the environment, - detecting a first element on the received image, - determining a first visual signature for detection, - calculating a distance between the first visual signature and a reference visual signature, - comparing the calculated distance to a first threshold and a second threshold, - identifying the first element as being the reference element, when the compared distance is less than or equal to the first threshold, the first element being unidentified when the compared distance is strictly greater than the second threshold.