Fingerprint Authentication Using Multi-Layer Perceptron Neural Network

Security has been an important issue and concern in the information systems. There is possibility that illegal access to some restricted data or devices may happen. Authentication of users is the key factor in the security of information systems. User authentication mostly based on PIN, password, smart card or biometrics. Recently biometrics based authentication like fingerprint is widely used to identify legitimate users. Biometrics authentication has an advantage over other methods in that it is based on something you are, which is not easily forgotten, copied or stolen. In this article, Multi-Layer Perceptron (MLP) neural network is trained with back-propagation (BP) algorithm to build the trained MLP neural network as matcher module. A feature extractor finds minutia features such as ridge end and bifurcation, from the input fingerprint images. The digital values (sum of coordinates) of these features are applied to input of the neural network for training purpose. For user authentication, the verification part of the system identifies the fingerprint based training performance of the network. Our proposed method is useful in solving the security problems that occurred in other fingerprint authentication systems, such as systems using the verification table

Keywords:

Authentication , biometric , fingerprint , Neural Network , MLP