Integration of Fingerprint Centre Point Location and Principal Component Analysis for Fingerprint Verification

This paper presents an efficient algoritlun for estimating the location of the centre point (CP) of a fingerprint for the purpose of aligning fmgerprints in a fingerprint verification system. Principal Component Analysis (PCA) is utilized in this system. The proposed CP estimation algoritlunis based on the Alteration Tracking (AT) followed by an estimation algorithm. AT is proposed toextract a track that records the transition from one quantized direction to another while the Centre Point Estimation (CPE) algoritlun is used to find the bending point with highest direction transition from the transition track. Fingerprint feature extraction is obtained with respect to the CP of the fingerprintdirectional field. PCA is then performed over the centred fingerprint directional field to extract importantinformation from fingerprint features for verification purpose. Experimental result shows that theproposed CP algoritlun is capable of reliably aligning fingerprints. Furthermore, PCA is able to extractinformation from aligned fingerprint directional field for fmgerprint verification.