ELECTRICA

A CNN Based Rotation Invariant Fingerprint Recognition System

1.

Department of Electric-Electronic Engineering, Faculty of Engineering, Osmangazi University, Eskişehir, Turkey

2.

Department of Electrical-Electronic Engineering, Faculty of Engineering, İstanbul Kültür University, İstanbul, Turkey

ELECTRICA 2017; 17: 3471-3479
Read: 1004 Downloads: 586 Published: 27 July 2017

This paper presents a CellularNeural Networks (CNN) based rotation invariant fingerprint recognition systemby keeping the hardware implementability in mind. Core point was used as areference point and detection of the core point was implemented in the CNNframework. Proposed system consists of four stages: preprocessing, featureextraction, false feature elimination and matching. Preprocessing enhances theinput fingerprint image. Feature extraction creates rotation invariant featuresby using core point as a reference point. False feature elimination increasesthe system performance by removing the false minutiae points. Matching stagecompares extracted features and creates a matching score. Recognitionperformance of the proposed system has been tested by using high resolutionPolyU HRF DBII database. The results are very encouraging for implementing aCNN based fully automatic rotation invariant fingerprint recognition system.

Files
EISSN 2619-9831