Alexandre Heili awarded the EPFL PhD degree for his work on "Human Tracking and Pose Estimation in Open Spaces"
In his research, and in the context of the European project VANAHEIM, Alexandre Heili developed perceptual algorithms for the estimation of behavioral cues, with the motivation that improving their output could in turn benefit human behavior analysis at a higher level. Location-based analysis being a first step towards making surveillance systems intelligent, Alexandre first designed a novel, efficient framework for multi-person tracking, enabling to consistently and simultaneously follow multiple people over time. Second, as pose cues like body and head orientations are able to characterize people's activities and interactions more precisely, he proposed robust methods for their joint estimation.
To download his thesis, click on the following link: Human Tracking and Pose Estimation in Open Spaces