Improved Liveness Testing Techniques for Biometric Applications

The subject of this proposal is the development of an optical sensor and algorithms for performing fingerprint spoof detection. The goal is to develop a compact, rugged, integrated optical sensor that collects both a conventional fingerprint image as well as a multispectral image of the subsurface skin characteristics. The multispectral image will be processed to ensure that the optical properties match those expected, and thereby confirm that the fingerprint was collected from a real, live, unaltered finger. The collection of both images will occur with a single user action, which has the advantages of greater user convenience, high user throughput, and increased security by having the biometric acquisition and spoof detection occurring on the same skin site at the same point in time. Once developed, the sensor and algorithms will be evaluated for vulnerabilities by an independent testing group. The results of this evaluation will be incorporated into a redesign of the technology that will then be the basis for the final deliverable from this project.