Fingerprint verification system using principal component transformation eigenvectors technique

Abstract EN

A fingerprint verification technique is presented.

The Principal Component Transformation “PCT” is used to identify a checked person, by comparing its eigenvector with the PCT eigenvectors of stored fingerprints images in a Database.

The computed Eigenvectors of the input images are determined from the covariance matrix of the set of fingerprints reduced mean images which were ascended as column vectors in a two dimensional array.

The covariance matrix then determined by multiplying the mean (i.e.

average) reduced matrix by its transpose.

The similarity between the stored images of the fingers (as Database) and the test finger is presented by utilizing the Minimum Mean Square Error (MMSE) criterion.