Comparison between proposed Convolutional Neural Network and KNN For Finger Vein and Palm Print
Hiba Wasmi, Mustafa M. Alrifaee, Ahmad Al Thunibat, Bassam M Al-Mahadeen · 2021
Biometrics technologies achieved an important role in facilitating the identification process of persons and accessing the secure areas in its success levels rather than the traditional methods, such cards, passwords, etc. In this research, we designed a multi biometrics recognition and authentication system using a proposed deep learning algorithm, called convolutional neural network, that depends on finger vein and palm print to treat the shortcomings like contrast of light, time complexity and accuracy. Histogram equalization was used for enhancement of image. Moreover, to extract the best feature, Linear Discriminate Analysis (LDA) method was applied by reducing the redundant features. The result of the proposed system gave 99%, which is higher than using of KNN technique.