Illumination alignment using lighting ratio: Application to 3D-2D face recognition

Illumination alignment is accomplished by relighting or unlighting methods that compensate the appearance differences due to variations in illumination conditions between a pair of face images or among a set of face images. In this paper, we propose a lighting ratio based method that avoids parameter tuning process by using an image-specific low-pass filtering. In addition, it includes an optimization step that brings the images into a similar energy level. The proposed method is simple and does not require any prior knowledge of the number, position, type of lighting sources or the 3D face geometry. Experimental results indicate that our method is effective and efficient for illumination-robust face recognition.