Trimodal Biometric Authentication System using Cascaded Link-based Feed forward Neural Network [CLBFFNN]

Abstract

The beginning of the 21st century was rich in events that turned the world’s attention to public security. Increase in technological advancement gave people possibilities of information transfer and ease of physical mobility unseen before. With those possibilities comes risk of fraud, theft of personal data, or even theft of identity. One of the ways to prevent this is through biometric authentication system. This paper considers a multi-biometric system involving a combination of three biometric traits: iris, fingerprint and face in order to make authentication cheaper and more reliable. When the images are captured using optical scanner and webcam, image pre-processing is done using Enhanced Extracted Face (EEF), Plainarized Region of Interest (PROI) and Advanced Processed Iris Code (APIC) methods for face, fingerprint and iris images respectively. These are fed into a Cascaded Link-Based Feed Forward Neural Network (CLBFFNN) which is a classifier trained with back-propagation algorithm. CLBFFNN comprises of CLBFFNN(1) used for training and CLBFFNN(2) used as the main classifier. Fusion of outputs from face, fingerprint and iris recognition systems is done at decision level using AND operation. With the use of the improved preprocessing methods, Optical Character Recognition (OCR) with intelligent barcode and CLBFFNN, the proposed intelligent multibiometric system is proved to be cheaper, more secure and efficient than the existing methods.