A Low-Cost Optical Sensor for Secured Antispoof Touchless Palm Print Biometry

Conventional biometric spoof detection sensors suffer from several problems, including complex expensive sensor assembly and design. They are susceptible to environmental perturbation and are not robust against all types of spoof attacks. Further, for a secured biometric sensor, the hierarchy of spoofing attempts (i.e., curious attempts, low-level attack, and high-level attack) needs to be precisely distinguished to perform different actions (e.g., a high-level attack should be handled carefully). In this article, we propose a novel three-level antispoof palm print biometric sensor using low-cost components and simple optical techniques (photography, fringe projection, and biospeckle analysis). At the first level, to eliminate the curious futile users, a 2-D image of the palm print is captured, and feature matching is performed. Next, the most prevalent low-level attacks (photo and layered attacks) are detected using fringe projection profilometry. At the final level, a novel biospeckle index based on a subtraction average technique is introduced to quantify the liveliness map for eliminating the high-level attacks, including fake prints and cadavers. The robustness of the biospeckle analysis against different possibilities to generate false activity (tremor, air flow, and heat flow) has also been verified.