A fast image enhancement algorithm for fingerprint

The reduction of time cost is a challenging task in an embedded fingerprint verification system. This paper proposes a fast image enhancement algorithm for fingerprint verification. Firstly, a two-level convolution template is designed according to ridge direction and ridge circle. Then a simplified convolution is used to enhance fingerprint, which speed up the enhancement procedure. Experimental results show that in general the proposed algorithm has similar or better enhancement results than the well-known Gabor enhancement. Moreover, the time cost is much less than that of Gabor enhancement. A test on a fixed-point DSP shows that the proposed algorithm is suitable for practical application in embedded mode.