Thesis of Romain Guesdon

Recognition of human behaviors by hybrid deep learning: an application to passengers of autonomous vehicles


Within the context of the region project AutoBehave, the main objectives of this thesis are to study and develop new computer vision and artificial intelligence techniques applied to the automatic recognition of human behaviour in confined environments such as vehicle interiors, at several scales in space and time.

Advisor: Laure Tougne
Coadvisor: Carlos Crispim-Junior