Experimental study of thumbprint-based authentication framework for ATM machines

The operation of Personal Identification Number (PIN) based Automated Teller Machine (ATM) has continued to experienced challenges currently militating against its acceptance and patronage. The challenges include card swallowing, misplacement, damage or expiration. Several users are also experiencing forgetfulness of PIN and the activities of fraudsters targeted at PIN pilfering for authorized access to account. This paper presents the experimental study of a framework for fingerprint authenticated ATM that addresses these challenges. The framework is an embodiment of different components for fingerprint processing which include enrollment, database and verification. Fingerprint enhancement, feature extraction and matching sub-modules form the verification component. The matching sub-modules rely on suitable mathematical models to function. The implementation was carried out using Windows 7 as operational platform while C# and Microsoft SQL server served as the frontend and backend engines respectively. The experimental study of the framework was performed with FVC2004, FVC2006 and FVC2008 standard databases and False Rejection Rate (FRR), False Acceptance Rate (FAR) and Average Matching Time (AMT) metrics were used for the evaluation. Obtained results were compared with those from similar and recently formulated algorithms.