The Study, Based On The Pre-processing Algorithm Of The Fingerprint Curve Characteristics

Automated Fingerprint Identification technology is one of the biometric Identification technologies. It depend on the information contain in the fingerprint skin to check identity of mastership. Because of the uniqueness and immutability of fingerprint, along with real time and practicability of the identification technique, fingerprint identification become to be one of the most popular, convenience and reliable methods for identification.Fingerprint Identification process include four important steps: 1 image acquisition, 2 image pre-processing, 3 feature extraction, 4 minutia matching. In our text we make research and discussion on the pre-processing of image: image enhancement and thinning. This paper introduces an enhancement and thinning approach based on cross oriented filter. Based on the Orientation estimation of image, to segment the legible part and blurry part, and then use the high-pass to enhance the high-frequency part, low pass filter to filtrate high-frequency part to wipe out the noises. We introduce the local adapted methods for binarization. We designed an edge-cut thinning template which reserving ridges structure of fingerprint and protecting the true minutiae.In order to test the performance of the approach, we have make experiment on fingerprint images provided by FVC2000 and 2004 databases, implement on PC by VC++. The experimental results demonstrate that it enhance the image effectively, reserves texture information and the true minutiae of fingerprint The result image laid a better foundation for feature extraction and matching.