Fake Face Recognition using Fusion of Thermal Imaging and Skin Elasticity

Spoofing with photograph or video is one of the most common manners to attack a face recognition system. In this paper, we present a non intrusive and real time method to address this problem, based on fusion of thermal imaging and skin elasticity of human face. In this technique, face images is captured using camera sensor and thermal sensor at the same time. Before capturing the images, user is asked to do some movement like chewing and forehead movement simultaneously, so that a full movement to face skin can be given and then sequence of face images is captured with a gap of few milliseconds. Then one of the image from image sequence is compared with the thermal image after thinning and noise removal. Then after applying correlation coefficient between images and then discriminate analysis using some method, face skin is discriminate from the other materials like gelatin, rubber, cadaver, clay etc. In comparison to other face liveness detection, this method will be much user friendly. On the other hand, one of the images captured for liveness detection can be used for face recognition.