Last Update: 2017-05-03
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 (Top View Person Re-identification) dataset stores depth frames (640x480) 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 100 people of TVPR were acquired in 23 registration session. The recording time [s] for the session and the number of persons of that session are reported in the following table. Each of the 23 folders contains the video of one registration sessions. Acquisitions have been performed in 8 days and the total recording time is about 2000 seconds. 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: 43 females and 57 male; 86 with dark hair, 12 with light hair and 2 are hairless. Furthermore, of these people 55 have short hair, 43 have long hair. The subjects were recorded in their everyday clothing like t-shirts/sweatshirts/shirts, loose-fitting trousers, coats, scarves and hats. In particular, 18 subjects wore coats and 7 subjects wore scarves. All videos have fixed dimensions (640x480 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.
| Video ID | time[s] | # person |
|---|---|---|
| g001 | 68.765 | 4 |
| g002 | 53.253 | 3 |
| g003 | 50.968 | 2 |
| g004 | 59.551 | 3 |
| g005 | 75.571 | 4 |
| g006 | 128.827 | 7 |
| g007 | 125.044 | 6 |
| g008 | 75.972 | 3 |
| g009 | 94.336 | 4 |
| g010 | 116.861 | 6 |
| g011 | 101.614 | 5 |
| g012 | 155.338 | 7 |
| g013 | 102.283 | 6 |
| g014 | 92.028 | 5 |
| g015 | 126.446 | 6 |
| g016 | 86.197 | 4 |
| g017 | 95.817 | 5 |
| g018 | 57.903 | 3 |
| g019 | 82.908 | 5 |
| g020 | 87.228 | 4 |
| g021 | 42.624 | 2 |
| g022 | 68.394 | 3 |
| g023 | 56.966 | 3 |
| Total | 2004.894 | 100 |
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 form: request form
- Send it to: vrai@dii.univpm.it (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:
@Inbook{Liciotti2017reid,
author={Liciotti, Daniele and Paolanti, Marina and Frontoni, Emanuele and Mancini, Adriano and Zingaretti, Primo},
editor={Nasrollahi, Kamal and Distante, Cosimo and Hua, Gang and Cavallaro, Andrea and Moeslund, Thomas B. and Battiato, Sebastiano and Ji, Qiang},
title={Person Re-identification Dataset with RGB-D Camera in a Top-View Configuration},
bookTitle={Video Analytics. Face and Facial Expression Recognition and Audience Measurement: Third International Workshop, VAAM 2016, and Second International Workshop, FFER 2016, Cancun, Mexico, December 4, 2016, Revised Selected Papers},
year={2017},
publisher={Springer International Publishing},
address={Cham},
pages={1--11},
isbn={978-3-319-56687-0},
doi={10.1007/978-3-319-56687-0_1},
url={http://dx.doi.org/10.1007/978-3-319-56687-0_1}
}
TVPR Dataset Results