A real time decision support system using head nod and shake

Abstract

Human gestures are very prominent means to interface with intelligent systems naturally and nonverbally. This paper presents a robust and real time vision based decision support system that automatically predicts the head nod and shake gestures for decision making. Here we apply the Gentle Adaboost algorithm that detects the faces in the video frames automatically. Real time eye tracking is performed to find the center coordinates of the eyes from the detected face region. In this work, we propose support vector machine classifier, a robust algorithm based on machine learning for predicting head nod and shake gestures. Moreover, this system can function as real time decision support tool in an online environment. The proposed system is implemented and tested for several real time videos. The experimental results show that this system is able to detect head nod and shake gestures with a detection rate of 91.1%.