TVPR\2 Dataset

Person re-identification is an important topic in scene monitoring, human computer interaction, retail, people counting, ambient assisted living and many other computer vision research. The TVPR\2 (Top View Person Re-identification - 2nd edition) dataset stores depth frames (320x240) collected using Asus Xtion Pro Live in top-view configuration. This setup choice is primarily due to the reduction of occlusions and it has also the advantage of being privacy preserving, because faces are not recorded by the camera. The use of an RGB-D camera allows to extract anthropometric features for the recognition of people passing under the camera.

Dataset structure, download and conditions of use

The 1000 people of TVPR\2 were acquired in more than 200 registration session. The recording time [s] for the session, the number of persons of that session and the passing order has been manually labelled and is available.. Acquisitions have been performed in 8 days. Registrations are made in an indoor scenario, where people pass under the camera installed on the ceiling. Another big issue is environmental illumination. In each recording session, the illumination condition is not constant, because it varies in function of the different hours of the day and it also depends on natural illumination due to weather conditions. 

The recruited people are aged between 19 - 36 years and belogns to different ethnic groups. The subjects were recorded in their everyday clothing like t-shirts/sweatshirts/shirts, loose-fitting trousers, coats, scarves and hats. Three couple of twins are also present. All videos have fixed dimensions (320x240 pixels) and a frame rate of about 30fps.

Videos are saved in native .oni files (depth and color streams), but can be converted in any other format. Colour stream is available in a non compressed format.

To obtain this dataset, we ask you to complete, sign and return the form below. After that, I will send you the credentials to download it. Note that the dataset is available only for research purposes.

  • Fill out this formrequest form
  • Send it to: (Note: you should send the email from an email address that is linked to your research institution/university)
  • Wait for the credentials
  • You will be sent a link for the download.


Please cite our work using the following bib:

author={Paolanti, M. and Pietrini, R. and Mancini, A. and Frontoni, E. and Zingaretti, P.},
title={Deep understanding of shopper behaviours and interactions using RGB-D vision},
journal={Machine Vision and Applications},



TVPR Dataset/2 Sample