Hand And Finger Gesture Recognition System for Robotic Application

Abstract

Vision-based gesture recognition system is the attractive solution for human computer interaction and machine vision application like robotic application. The common systems assumption under such a systems are a constant environment, like persons wearing non-skin-colored clothes with long sleeves and a fixed camera position under constant lighting conditions. In this paper we are evaluating the performance of a simple hand and figure gesture recognition system that can be mapped to various tasks for robotic application. One of the main constrain for implementing the vision based gesture recognition system on robotic platform is the computational complexity associated with the system because of the low computation power available on the tiny embedded system on robots. So here we are evaluating a simple hand gestures for that purpose based on a simple heuristic algorithm.