Welcome to Mirror Worlds Challenge

Mirror Worlds is an NSF funded infrastructure project at Virginia Tech ICAT, focused on creating a computing infrastructure for research in mixed reality and smart environment. More details here. An important part of this infrastructure is to use 35 ceiling mount fisheye cameras in Moss Art Center building at Virginia Tech, Blacksburg VA to detect and track people in real time. This is a very challenging topic in computer vision research, so we created this dataset.

Mirror Worlds challenge is a Multi-Target, Multi-Camera Tracking dataset. It involves several computer vision research topics including detection, tracking, person re-id. All under real time constraint and in real world environment.

  • 35 ceiling mount fisheye cameras with known geometry.
  • Real world environment with occlusions and diverse appearance.
  • Detection and tracking in real time.
  • Both real and synthetic dataset.

Dataset Example

What's new

  • Mar1, 2018: We are releasing some preliminary video and results. We plan to release our pilot dataset in less than a month.
  • Apr 1, 2017: Our initial challenge website is online: here
  • Jan 27, 2017: Our Mirror Worlds website is online: here


This dataset is for academic research use only. However, we take your privacy seriously! If you find yourself or personal belongings in this dataset and feel unwell about it, please contact us and we will immediately remove the respective data from our server. Most of our data is collected with IRB approved studies and volunteers.


This research is supported by the National Science Foundation under Grant No. 1305231.

Creative Commons License by Nuo Ma

Design courtesy : motchallenge.net & Apple Inc.