Skeleton based People Re-Identification for Surveillance Systems in Medical Environments



  • Fall 2016-present

Project Description

This project is a part of the Tustworthy Health & Wellness (THaW) project, which is an NSF-funded project that tackles many of the research challenges to provide trustworthy information systems for health and wellness.

Currently, surveillance systems have been deployed in many medical environments to provide security assurance, e.g. tracking Alzheimer’s patients to avoid wandering. However, currently the surveillance camera based people re-identification problem is still challenging because of diverse factors such as various poses and light conditions. This project aims to design a robust people re-identification method based on skeleton information extracted from 2D images. We plan to use a deep neural network to locate fixed positions on a human body, and design local-comparison based features to rule out the impacts of human poses and light conditions. The results of this project can be used to enhance the automatic human tracking systems for medical environment surveillance.


  • Tuo Yu, Haiming Jin, Wai-Tian Tan, and Klara Nahrstedt, “SKEPRID: Pose and Illumination Change-Resistant Skeleton-Based Person Re-Identification,” in ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), vol. 14(4), pp. 1-24, 2018.

Funding Agencies

This project is supported by the a collaborative award from the National Science Foundation (NSF award numbers CNS-1329686, 1329737,1330142 and 1330491).