Recognizing human interactions from video is an important step forward towards the long-term goal of performing scene understanding fully automatically. Recent years have seen a concentration of works revolving around the problem of recognizing single person actions, as well as group activities, while the area of modeling the interactions between two people is still relatively unexplored. In (Motiian et al., 2013) people interactions are modeled by forming temporal interaction trajectories coupling together the body motion of each individual as well as their proximity relationships. Such trajectories live in a well-defined Riemannian manifold and enjoy specific symmetry properties that have to be taken into account during the development of a theoretically grounded recognition framework.
- ISVCPairwise Kernels for Human Interaction Recognition In Advances in Visual Computing, 2013. Oral