I am a PhD student in the Perception team since September 2017, supervised jointly by Radu Horaud and Xavier Alameda-Pineda.
I received an engineering degree in applied mathematics and computer science from the Ensimag Institut polytechnique de Grenoble in 2017.
I work on the applications of machine learning in computer vision, and am especially interested in Deep Learning, Domain Adaptation, Appearance Modeling, Person re-Identification and visual tracking.
Thesis Title: Unsupervised domain adaptive multiple person tracking and visual identification for human-robot interaction
Human robot interaction requires the robot to have an accurate knowledge of its environment, especially who is present, and where, to enable an interactive conversation. In this context, my thesis proposes to exploit image information recorded by the embedded camera to perform Multiple Object Tracking (MOT), leveraging localization and identification by exploiting temporal and spatial proximity to produce ID-exploitable trajectories.
State-of-the-art methods rely on deep learning approaches, which are known to heavily depend on the training data, and suffer from poor generalization ability. More specifically, most of MOT implementations embed a person re-identification model to use as appearance cue, while those are widely known to be sensitive to background changes and illumination conditions. Consequently, my thesis focuses on investigating adaptation strategies to new domains for MOT and re-ID models. The PhD manuscript detailing those approaches will be publicly available after the defence.
Contact
- INRIA Grenoble Rhone-Alpes
655, avenue de l’Europe
38330 Montbonnot Saint-Martin
France - Email: guillaume dot delorme at inria dot fr
guillaume0delorme at gmail dot com
