The performance of a fingerprint recognition system relies on the quality of the input fingerprint images. Several researches have been studied on the enhancement of fingerprint images for fingerprint recognition. The representative enhancement is the adaptive filtering method based on Gabor filter (GF). However, this method is computationally expensive due to the large mask size of GF. In this paper, we propose a half Gabor filter (HGF), which is suitable for fast implementation in spatial domain. The HGF is a modified filter which preserves the frequency property of a GF and reduces the mask size of the GF. Compared with the GF, the HGF not only reduces the processing time approximately by 41% but also enhances the fingerprint image which is as reliable as the GF.

No abstract