Human-robot communication through visual game and gesture learning

摘要

Human Machine Interaction and Gesture Recognition is a very challenging field. Scientists and researchers are trying to make communications with machines as easy, smooth and reliable as possible. With the help of learning techniques we present a vision based simulation of CHALK, BLACKBOARD, DUSTER game in which Robot plays this game with human identifying the moves generated by him in the form of gestures and also learns new gestures. Also with presented approach we are able to perform forearm deletion. We have verified our work using Webots simulation platform.