Enhancing Biometric Authentication with Convolutional Neural Networks for Finger Vein and Palm Recognition

This study explores the enhancement of biometric authentication using Convolutional Neural Networks (CNNs) for finger vein and palm recognition. For palm recognition, a dataset of palm images is utilized, followed by pre-processing steps such as image resizing. CNN classification is then applied to classify the palm images, with the ultimate goal of achieving high accuracy in authentication. Similarly, for vein recognition, a dataset of vein images is employed, along with pre-processing techniques including image resizing, restoration, and noise removal. CNN classification is then utilized for vein image classification, aiming to achieve accurate authentication. If both palm and vein images belong to the same person, authentication is successful; otherwise, it is deemed unsuccessful. This approach combines the unique characteristics of both palm and vein biometrics to enhance the security and accuracy of biometric authentication systems. The use of CNNs allows for effective feature extraction and classification, enabling robust authentication based on palm and vein patterns. Overall, this study presents a promising approach to advancing biometric authentication systems by leveraging CNNs for palm and vein recognition.

Keywords: Finger Vein and Palm Dataset, Deep Learning, CNN, Preprocessing, Classification and accuracy.