Automated Contact Tracing Using Person Tracking and Re-Identification

Conference: IEEE 2nd International Conference on Secure Cyber Computing and Communications (ICSCCC), Jalandhar, India

Authors: Abhay Khandelwal, Aditya Kotwal, Pratham Sutone, Vedant Wagh

This paper introduces a novel contact tracing system that automates the identification of potential infection risks using computer vision and deep learning techniques. The system leverages CCTV footage to detect individuals, track movements, and analyze interactions for generating a contact tracing graph.


Abstract

The proposed system automates the process of manual contact tracing by:

  • Detecting and identifying individuals in surveillance footage.
  • Tracking interactions and estimating distances between individuals.
  • Storing, filtering, and analyzing the data to provide actionable insights.

This approach is designed to improve the efficiency and scalability of traditional contact tracing methods, especially during pandemics such as COVID-19.


Highlights

  1. Real-Time Detection:
    • Implemented person tracking using YOLOv3 and Kalman filters.
  2. Depth Estimation:
    • Calculated spatial distance using triangle similarity techniques for accurate contact detection.
  3. Contact Tracing Graph:
    • Created a graph-based representation of interactions to identify individuals at risk.
  4. Performance Metrics:
    • Achieved 91% accuracy based on confusion matrix analysis.

Citation

Khandelwal, A., Kotwal, A., Sutone, P., & Wagh, V. (2021). “Automated Contact Tracing using Person Tracking and Re-Identification.” IEEE International Conference on Secure Cyber Computing and Communications (ICSCCC), pp. 102-107. DOI: 10.1109/ICSCCC51823.2021.9478143


Explore the full paper on IEEE Xplore