Low Cost Activity Recognition Using Depth Cameras and Context Dependent Spatial Regions

Michael Karg (1) , Alexandra Kirsch (2, 1, 3)

1 Institute for Advanced Studies
2 Intelligent Autonomous Systems
3 Human-Computer Interaction and Artificial Intelligence

Recognition of human activities is usually based on expensive sensor setups to extract rich information such as body posture or object interaction. We investigate the use of inexpensive depth cameras to perform activity recognition using context dependent spatial regions with two different approaches: Spatio-Temporal Plan Representations and Hierarchical Hidden Markov Models. We evaluate both approaches in a simulated and a real-world environment.

Mots clés: Human Factors; Activity Recognition; Human Robot Interaction

Domaines: Intelligence artificielle [cs.AI]; Robotique [cs.RO]; Interface homme-machine [cs.HC]

Michael Karg, Alexandra Kirsch. Low Cost Activity Recognition Using Depth Cameras and Context Dependent Spatial Regions. 13th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2014), 2014, Paris, France. ⟨hal-01691651⟩