Invision AI develops a multi-camera system to track people and cars. The current system sometimes mixes up persons in close proximity. The goal of this project is to explore the best ways of using people’s appearance to avoid such confusion.
Led by the recent popularity of the person re-identification task, research has made great strides in appearance-based tracking. The student will start by exploring recent state-of-the-art methods. Once familiar with the field, the most appropriate model will be applied on real data acquired by Invision AI and combined with their existing tracking framework.
Once working as desired, the approach will be extended to deal with more challenging scenarios, such as multi-camera setup, crowded scene, partial occlusion and/or significant illumination changes.
Questions and contact
[email protected] [email protected]
Supervision
This project is co-supervised by Julien Pilet and Carlos Becker at Invision AI, in Renens, and by Martin Engilberge at CVLab, EPFL.
Rules
https://www.epfl.ch/schools/ic/education/master/master-project/
When
25 weeks, starting in mid/end Feb 2023.
Where
At Invision AI office, rue de Lausanne 64, 1020 Renens and CVLab, EPFL