Gesture based imitation learning for Human Robot Interaction

This paper describes a gesture based recognition system using Indian Sign Language (ISL) for performing Human Robot Interaction (HRI) in real time. It permits us to construct a convenient gesture based communication with humanoid robot HOAP-2. The classification process is carried out by extracting the features from ISL gestures. Orientation Histogram is considered as a feature vector for classification process. It is to be done by the two statistical approaches namely known as Hidden Markov Model (HMM) technique and Bhattacharyya Distance estimation in order to achieve satisfactory recognition accuracy. The major task involves by computing the recognition time taken by both the above methods and finally the suitable one participates for HRI applications. The classification result has been tested on the Webots simulation platform on Humanoid robot (HOAP2) to generate mimicry according to the recognized