Detection of Conversational Groups and Tracking of Lower Body Orientations

Appropriate robot behavior in public, open spaces cannot occur without the ability to automatically detect conversational groups of free-standing people. To this end, we proposed an alternating optimization procedure that estimates lower body orientations and detects groups of interacting people. The first task was achieved by tracking the direction of the lower body of the people in the scene based on their position, their head orientation, the location of objects of interest in their vicinity, and their groups. For the second task, we proposed a new group detection algorithm based on F-formation detection. This method can reason about lower body orientation distributions, and generates soft group assignments for the orientation trackers. We evaluated the proposed approach on a publicly available dataset, and showed that it can improve state-of-the-art detection of non-interacting people without sacrificing group detection accuracy. This is particularly useful for robots since it provides more opportunities for starting interactions and can help estimate disengagement.


We thank the Walt Disney Corporation for their support. We also thank O. Lanz for the Cocktail Party dataset, M. Cristani for their group detection code, and E. J. Carter for her assistance on this project.

M. Vázquez, A. Steinfeld, S. E. Hudson. Parallel Detection of Conversational Groups of Free-Standing People and Tracking of their Lower-Body Orientation. Proc. of the 2015 IEEE International Conference on Intelligent Robots and Systems (IROS), 2015.