AdMiRe-EPFL dataset

The AdMiRe-EPFL dataset is a series of videos recorded at the MMSPG laboratory for its usage in the H2020 AdMiRe project. It contains scenes with different subjects in the foreground. The scenes were recorded with different devices like webcams and smartphones close to the use case of the project. They were used to qualitatively evaluate the results of different background extraction methods.


The dataset is composed of 13 raw video shot by consumer devices and 33 processed videos. The videos contain different scenes of subjects in front of the camera with different backgrounds (including a physical green screen), and different configurations (full body shots, conference-like head and shoulder shots…).

Conditions of use

In no event shall the Ecole Polytechnique Fédérale de Lausanne (EPFL) be liable to any party for direct, indirect, special, incidental, or consequential damages arising out of the use of the data and its documentation. The Ecole Polytechnique Fédérale de Lausanne (EPFL) specifically disclaims any warranties. The data provided hereunder is on an “as is” basis and the Ecole Polytechnique Fédérale de Lausanne (EPFL) has no obligation to provide maintenance, support, updates, enhancements, or modifications.

In case you use this dataset in your research we kindly ask you to cite the paper:

[1] H. Piñeiro Monteagudo, R. Daod Nathoo, L. Deillon, C. Gao and T. Ebrahimi, “AI-based telepresence for broadcast applications,” Proc. SPIE, Applications of Digital Image Processing XLV.


This dataset was generated with contributions from the H2020 Innovation Action “Advanced Mixed Realities (AdMiRe)” under grant agreement 952027.


The dataset can be downloaded as a zip file from the following FTP by using dedicated FTP clients, such as FileZilla or FireFTP (we recommend to use FileZilla):

Protocol: FTP
FTP address:
Username: [email protected]
Password: ohsh9jah4T
FTP port: 21

After you connect, choose the AdMiRe-EPFL folder from the remote site, and download the relevant material.

Please refer to the README file for further information on the structure and the usage of the material.