Robust Fingerprint Classification Using An Eigen Block Directional Approach

This paper describes a method of fingerprint classification using Eigen Block Directional Fingerprints. The method we propose dispenses off with the preprocessing stages such as segmentation, binarisation, thinning etc. We determine the block directional representation of a fingerprint and the location of the core points in a set of template fingerprints which belong to the same class. In the next step, we determine the most prominent eigen vectors of each template, which we term as the Eigen Block Directional Fingerprints. To determine the classification of a fingerprint, we extract the Block Directional Fingerprint of the query image and determine the alignment parameters between the template and the query images using eigen tracker which minimises the robust error. We declare the class of the query image as the class of the Eigen Block Directional Fingerprint which result in the least robust error norm.

• Type: Conference
• Year: 2004
• Where: ICVGIP
• Authors: P. Madhusoodhanan, Sumantra Dutta Roy
• Keywords: Block Directional Fingerprint, Eigen Block, Fingerprint, ICVGIP 2004, ICVGIP 2008