Infrared face recognition by using blood perfusion data

Abstract

This paper presents a blood perfusion model of human faces based on thermodynamics and thermal physiology. The target is to convert the facial temperature data which are liable to ambient temperature into consistent blood perfusion data in order to improve the performance of infrared (IR) face recognition. Our large number of experiments has demonstrated that the blood perfusion data are less sensitive to ambient temperature if the human bodies are in steady state, and the real data testing demonstrated that the performance by means of blood perfusion data is significantly superior to that via temperature data in terms of recognition rate. © Springer-Verlag Berlin Heidelberg 2005.