An Improved Algorithm for Fingerprint Image Enhancement

T.-L. Huang, Che-Wei Liu, C.-Y. Li, Cheng-Wei Liu, and C.-C. Huang (Taiwan)


Fingerprint, Gabor, Orientation field


In general, the quality of a fingerprint image obtained from a scanner is better than the fingerprints acquired from a criminal locality. It will deeply influence the correctness of the minutiae extraction and the final recognition. In order to increase the successful rate of the recognition, the enhancement of the fingerprint image is obviously important. We propose a new method to enhance the fingerprint image. First, we use a median filter to reduce noise of fingerprint images and then estimate the orientation field. Next, we give a new algorithm to modify the wrong orientation field caused by noise. And then we filter out noise through Gabor filter according to the modified orientation field. Finally, we get a binary fingerprint image. After the enhancement, minutiae can be directly extracted from the binary fingerprint image. Finally, the experimental results demonstrate the effectiveness of the proposal method.

Important Links:

Go Back