Study of Visible Face Recognition Methods Applied to Infrared Spectrum

Abstract

The face recognition methods originally implemented in the visible spectrum show that the recognition rate will not be necessarily the same as in the infrared spectrum. Such is the case of many visible methods for instance Local Derivative Pattern (LDP), Local Binary Pattern (LBP), Histogram of Oriented Gradients (HOG), Gabor Jet Descriptor (GJD) which obtains high performance in visible spectrum but they have not been studied in thermal domain. For that reason the aim of this article is to show the performance of such methods in both spectra using a comparative analysis of current methods for face recognition. In addition, this work shows characteristics in each spectrum indicating their advantages as robustness to illumination and changes in expressions for infrared spectrum and weaknesses as the major use of the face when articulating gestures for the visible spectrum.