Weighted Vote Fusion in prototype random subspace for thermal to visible face recognition

Abstract

The human body, like all other objects with temperature above the absolute zero, emits electromagnetic wave. The emission of infrared electromagnetic wave from the human face produces thermal images. Thus thermal images can be formed even in dark conditions, in which the formation of the visible image is impossible. However, the majority of the stored images in the recognition systems are visible. Thus, matching the thermal probe and visible gallery images can solve the night time face recognition problem. On the other hand, because of the different formation mechanism of these two types of images, there are lots of challenges in the matching process. Prototype random subspace approach is one of the most successful methods in the area of thermal to visible face recognition. In this paper, we have revised the recognition step of prototype random subspace approach by proposing Weighted Vote Fusion scheme. The proposed strategy has been tested on an available data set and the results show about 9% of improvement in recognition rate, comparing to the original approach.