Infrared Face Recognition in Forensics via Texture Analysis

Abstract

Bad lighting was one of the difficulties that facial-recognition systems had to overcome as the large intra-class variations due to pose, lighting, and expressions. Recently, there are many works using thermal cameras in face recognition systems. This paper recommends a thermal face recognition method based on texture analyses in a complex database including different poses and expressions. Three different texture analyses methods are compared in this problem. Due to the pose variations in the database Scale Invariant Feature Transform algorithm is more successful to other methods explained in the paper.