| The conventional personal identification methods, e.g. personal password, identity card, and key, are not satisfied to the need of modern society in security and convenience. As an identity verification method, fingerprint verification is one of the most important sub fields of biometrics. Based on fix-point DSP, fix-point operation is faster than floating-point operation. So how to improve the speed of fingerprint verification and reduce the error rate of fingerprint pre-processing are analyzed and researched.In an automatic fingerprint authentication system based on fix-point DSP, we get image of a finger in real-time by optics sensor. But the optic distorted fingerprint will increase the error rate of fingerprint feature matching. This paper proposes an effective geometrical transform algorithm to reduce the error rate of fingerprint feature matching and an optimizing algorithm of geometrical transform to save time cost of fingerprint pre-processing.Fingerprint verification algorithm includes two parts: fingerprint feature extraction and feature matching. There are many floating-point functions in the process of fingerprint feature extraction and feature matching processes. A fast algorithm uses sub section lines to near some functions of arc tangent. In the minutiae matching stage, we improve the minutiae-matching algorithm to degrade the time cost.Experimental results show that the fast algorithms reduce the time cost of fingerprint pre-processing and feature matching. Geometrical transform algorithm increase recognition rate and speed of automatic fingerprint authentication system. It is valuable in both academic and real application. |
PDF Full Text Request