Gesture Recognition Using Commodity RGB-D Sensor for Imitation Learning Platform for Children with Autism

Abstract

Autism Spectrum Disorders (ASD) is characterized by profound impairments in social interaction and communication. Children with ASD have deficits in core areas of social interactions such as gesture imitation. Gesture imitation is one of the early developed social communication skills and is thought to be linked with concurrent as well as later complex social skills such as language development, play and joint attention skills. Thus early identification of such a deficit and providing appropriate intervention regarding gestural imitation skills are quite important. This work, which is a part of a larger study that aims at building a gesture imitation intervention platform for children with ASD, leverages the intrinsic interest of the children with ASD in robotic technology. In this paper, we discuss the part of the robot-mediated intervention system that deals with gesture recognition and present preliminary recognition results.