CNN and Genetic Algorithm for Finger Vein Recognition

Abstract

Biometrics are modernized methodologies for checking or seeing the character of a living individual dependent on some physiological characteristics, like finger vein, palm vein, and iris or depending on behavioral characteristics like keyboard typing style, signature, and voice. Classical Finger Vein Recognition systems performed by identification depended on finger-vein lines elicited from the images, which were inputted or image improvement, and texture specification elicitation from the finger-vein images. This paper will study applying the Genetic Algorithm with Convolution Neural Network for finger vein classifying recognition system. The Genetic Algorithm's (GA) global searching ability is utilized to start the training process of a traditional Neural Network (CNN). Before training, the weights of the network are established using the GA genetic algorithm rather than random initializers. In terms of performance, the proposed strategy of employing a Genetic Algorithm and a Convolutional Neural Network (GA-CNN) performs better in terms of accuracy, sensitivity, and precision. GA-CNN is a new method and saves time in the training phase for detection and verification processes in biometric systems. This will be used in the future in the detection and face recognition systems.