Infrared-based face identification system via Multi-Thermal moment invariants distribution

This paper presents the implementation of moment invariants in an infrared-based face identification system to develop a novel and less complicated face identification system in thermal spectrum. Initially the system decomposes an infrared image into 4 thermal regions via 3-valued threshold method. In order to increase the information obtained from each decomposed IR image, the first four of the classical Hu´s moment invariants were computed from each generated thermal regions. Additional to this, centroid point and area (in pixel) were also obtained from each generated thermal region. The acquired Hu´s moment invariants together with the acquired centroid point and area (in pixel) forms a string of features for classification/identification. This system employs a hierarchical minimum distance measurement method for classification purposes. The performance of this system is encouraging with 87% of correct identification rate for test to registered image ratio of 2:1 and 84% of correct identification rate for test to registered image ratio of 4:1.