Onboard pedestrian tracking for driving assistance

Researcher: Damien Matti (PhD candidate)
Supervisor: Jean-Philippe Thiran (EPFL)

Description:

This research project will explore exploitation of behavior models in tracking systems has proven to be effective in coping with occlusion problems, using different types of on-board sensors, depending on their availability, can be used for detection.
Head pose of the pedestrian would be an additional input to predict his/her trajectory and behavior.