Person Re-ID

Mixed Precision Person Re-Identification for Edge Deployment

Person Re-ID implements a mixed precision approach for real-time person re-identification on edge devices.

Key Features

  • ResNet-50 and MobileNetV2 backbone support
  • NVIDIA Apex half-precision (FP16/FP32) inference
  • Real-time tracking on embedded GPUs
  • Trained on DukeMTC, CUHK03, and Market1501 datasets
  • Component of the REVAMP2T surveillance pipeline

Publication

M. Baharani, S. Mohan, and H. Tabkhi, “Real-Time Person Re-identification at the Edge,” ICIAR 2019. (17 citations)

Code