Incrementally verifying automata for fingerprint sensors

With the rapid development of fingerprint technology and computational algorithm, the accuracy requirements of semiconductor fingerprint sensors are getting higher and higher. From a business perspective or to a manufacturer, demands for automatic quality assessment and verification of semiconductor sensors are much critical. This paper has developed a complete evaluation system including extraction, matching algorithm, threshold regulation and cumulative similarity score. The SIN matching solution combined with single point and neighboring triple points mapping strategy is presented as a core of the matching algorithm. By means of repeated operations, an incremental verifying tool for estimating the quality of fingerprints is successfully built. Based on dozen sampling each user's fingertips, a Dozen Matching Module is built as a basic component in the system. Moreover, the method described in this paper can not only assess the full fingerprint or fragment fingerprint sensors, but also strengthen the system in the practical application of extensive and convenient.