An embedded fingerprint authentication system with reduced hardware resources requirement

In this paper, we present a fast, high accurate, and multiplierless fingerprint authentication method. Firstly, the fingerprint image is captured and performed noise canceling by a binary low-pass filter. Secondly, we estimate the displacement of the fingerprint pattern by means of directional image matching, and then circular profile features of the fingerprint are extracted for verification matching. Finally, the proposed method is implemented in an 8-bits microcontroller run at 25 MHz. The experiment results show that the proposed system possesses excellent recognition performance and efficiency. The advantage of the low complexity architecture is to apply fingerprint recognition in a low cost, resource-constrained embedded environment.