Neural Network based Fingerprint Classification

This paper introduces a new approach of fingerprint classification system based on ANN. Most automated fingerprint identification system uses prior classification of fingerprint using minutiae as feature. But in such methods the performance of minutiae extractions relies heavily on an enhancement algorithm and also it needs improvement as they are limited to the number of classable data. Thus many fingerprints are classified together, taking a long time to match and verify a given fingerprint. So instead of classification using minutiae we proposed a classification system that is based on individual features like singular point. Singular point detection is very robust and reliable, which overcomes the problem about rotation and translation. Classification efficiency is improved using Back Propagation Algorithm because we dont need to compare an input fingerprint image to all registered fingerprint images. The algorithm is tested accurately and reliably for many fingerprint images in FVC2004 database.