Curvelet feature based fingerprint recognition: Using fourier enhancement

Fingerprints are the oldest and most widely used form of biometric identification. Uniqueness of one´s Fingerprint remains same throughout the lifetime. Therefore, fingerprints are being used in many biometric systems such as civilian and commercial identification devices and forensic investigation systems. In our paper we have proposed a fingerprint recognition system as a combination of three significant stages: preprocessing, post-processing and matching stage that improves accuracy by overcoming the challenges previous systems presented. We have applied several enhancement methods in the pre-processing stage. For fingerprint image alignment purpose we have used core point based `Poincare Index value´ method, which is one of the widely used method for core point detection. Curvelet transformation, a new multi-resolution and multi-orientations tool is therefore used for the extraction of feature set from the image in different scale and orientation. Extracted features vectors from Curvelet co-efficient matrix are then classified using k nearest neighbor classifier in order to recognize the specific fingerprint.