Fingerprint image enhancement using Gabor wavelet transform

The performance of any fingerprint recognizer system highly depends on the fingerprint image quality; therefore fingerprint image enhancement is an essential stage in these systems. Wavelet transform is a good tool for non-stationary signal processing and has a wide range of application. In this paper a wavelet based method for fingerprint image enhancement has been proposed. In our algorithm Gabor wavelet and Gabor filter is used for this purpose. First the image is divided into blocks. 64 Gabor wavelet based on 4 frequencies and 16 directions are defined then the orientation and frequency of any image block are estimated. By using these estimations, a suitable Gabor filter is designed and then the image blocks are filtered. Comparing the proposed algorithm with other approaches shows improvement in the accuracy rate on a set of images from the FVC2002 database. Moreover, in the proposed method the frequency and direction are estimated simultaneously.