REVAMP2T

Real-Time Edge Video Analytics for Multi-Camera Pedestrian Tracking

REVAMP2T is an end-to-end scalable intelligent video surveillance system enabling privacy-aware multi-camera pedestrian tracking with real-time edge inference.

Key Features

  • Real-time multi-camera pedestrian tracking on NVIDIA Jetson Xavier
  • Privacy-aware analytics with on-device processing
  • Detection, re-identification, pose estimation, and segmentation models
  • 23 FPS for eight concurrent cameras at Full HD resolution

Publication

C. Neff, M. Mendieta, S. Mohan, M. Baharani, et al., “REVAMP2T: Real-Time Edge Video Analytics for Multicamera Privacy-Aware Pedestrian Tracking,” IEEE Internet of Things Journal, vol. 7, no. 4, pp. 2591-2602, 2019. (93 citations)

Paper Code