A Survey on Heterogeneous Face Matching: NIR Images to VIS Images

Human face matching plays an important role in applications such as criminal identification, banking, video surveillance and credit card verification. Face images are captured by different acquisition systems like visible light cameras and infrared imaging devices. Visible light cameras capture visible light images while near infrared images are captured by infrared imaging devices. These images are called heterogeneous images as they have different image formation characteristics like qualities of cameras, sensors used in cameras and different environment conditions. Matching heterogeneous images is subject of the research. This survey provides a review of different existing techniques for heterogeneous face matching. These techniques commonly categories into two parts namely Invariant feature based techniques and Common space based techniques. Moreover, the datasets available and performance parameters for evaluation are also discussed.