AdaBoost-based sensor fusion for credibility assessment

Individual credibility assessment is an ever-growing requirement in different applications including border crossing, airport checkpoint screening, criminal investigations, etc. In this paper, we describe a heterogeneous sensor fusion algorithm based on the AdaBoost algorithm and a speech emotional recognition technique in order to develop an automatic noninvasive credibility assessment system. Our initial results based on two available sensors show a promising increase in the accuracy and performance of the system.