Moments Invariant for Expression Invariant Thermal Human Recognition

Recent advancement of technologies in the field of bioinformatics and pattern recognition provides the opportunities for researchers and scientists to explore in depth the thermal human face image signals and to convert information into a meaningful knowledge through computational-based models, for the task of identification and recognition. Despite successes in indoor access control applications, imaging in the visible spectrum demonstrates difficulties in recognizing the faces in conditions of varying illumination and expression. Recently, researchers have investigated the use of thermal imagery for face recognition with good results. The use of thermal infrared images can improve the performance of face recognition in uncontrolled conditions. In this paper, we present a new technique for face recognition based on histogram distribution and moments invariants of thermal images of varying expressions. The new method has been tested on a new database comprising of images of different expressions and were taken within different time-lapse. Further, the database is challenged with images suffering from opaqueness. The experimental results have shown Rank-1 success rate of 98%.