Multi-DVS Circuitry for Privacy-Aware Person Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous stores face privacy concerns and regulatory limitations due to invasive image-based tracking, and dynamic-vision sensor (DVS) cameras offer a less invasive but more expensive alternative for event-based tracking.
Innovation Solution
A circuitry and method utilizing multiple DVS cameras for event-based tracking, recognizing individuals based on movement patterns across overlapping or non-overlapping fields-of-view, leveraging event-based visual data to track persons anonymously and efficiently, with the aid of Artificial Neural Networks for accurate identification and re-identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image-based tracking is used in autonomous stores, then tracking accuracy and visual information completeness are improved, but privacy invasiveness and regulatory compliance worsen
Solution Approach 1:
The patent extracts only the essential tracking information (movement patterns, position changes) from visual data while leaving out identifiable image information. DVS cameras capture only changes in pixel intensity over time, extracting motion events without capturing full visual frames, thus achieving tracking functionality while removing privacy-invasive visual details
Solution Approach 2:
Instead of capturing complete images and then analyzing movement, the system inverts the approach by capturing only movement events (changes in pixel state) and reconstructing tracking information from these minimal data points. This inversion fundamentally changes what is captured from the scene, prioritizing motion information over visual appearance
2Object-affected harmful factors
If DVS cameras are used for event-based tracking, then privacy protection is improved, but device cost worsens
Solution Approach 1:
The patent combines multiple DVS cameras to create a multi-camera event-based tracking system. By merging data from several cameras with overlapping or non-overlapping fields-of-view, the system achieves comprehensive coverage and reliable re-identification across camera boundaries, amortizing the cost of expensive DVS sensors across multiple tracking zones and improving overall system efficiency
3Reliability
If multiple DVS cameras are used for tracking across fields-of-view, then tracking reliability and re-identification capability are improved, but device complexity worsens
Solution Approach 1:
The patent segments the tracking space into multiple fields-of-view covered by different DVS cameras. Each camera independently tracks events within its own field-of-view, and the system manages multiple independent tracking streams. This segmentation allows parallel processing of visual data from different cameras, reducing the computational complexity of processing a single large-scale scene while improving tracking reliability through distributed observation
4Productivity
If event-based visual data is used instead of full images, then data processing speed is improved, but information completeness worsens
Solution Approach 1:
The patent employs dynamic event-based processing where the system continuously adapts to changing visual conditions by processing only relevant change events rather than static frames. The DVS cameras naturally adapt to varying illumination conditions by triggering events only when pixel intensity changes exceed a threshold, providing dynamic processing speed that maintains high productivity while preserving essential motion information across varying lighting conditions
Data Source
AI summary
The present disclosure pertains to a circuitry for event-based tracking configured to recognize a person based on event-based visual data from a first dynamic vision sensor camera and from a second dynamic vision sensor camera, and track the person based on a movement of the person when the person leaves a first field-of-view of the first dynamic vision sensor camera and enters a second field-of-view of the second dynamic vision sensor camera.