Optimization Models and Algorithms for Fingerprint Recognition and Its Applications in AFIS of China
摘要：In this talk, a general optimization model of automatic fingerprint identification system (AFIS) is proposed. To solve the general model, a serial of optimization models and algorithms are established and designed, including the modules of feature extraction and minutiae matching in AFIS. For the low-quality fingerprint images and the large database, we propose a global optimization model for orientation the field computation, a variable dimension optimization model for the singular point detection, and a bipartite graph optimization model for the minutiae matching. According to the characteristics of fingerprint image, corresponding novel algorithms are designed for these three models. These algorithms are embedded in our AFIS, which has been successfully applied to many provinces (cities) in China, and played an important role in cracking and preventing all kinds of criminal cases.