Open AccessConvolution Neural Network Based Deep Feature Fusion for Palmprint and Handvein
Author(s) -
Shreyas Rangappa,
Chigurupati Naveen,
H. K. Chethan,
K. S. Raghunandan,
C. N. Sunil
Publication year - 2019
Publication title -
international journal of engineering and advanced technology
Language(s) - English
Resource type - Journals
ISSN - 2249-8958
DOI - 10.35940/ijeat.f9217.109119
Subject(s) - convolution (computer science) , computer science , biometrics , artificial intelligence , convolutional neural network , feature (linguistics) , feature extraction , pattern recognition (psychology) , deep learning , artificial neural network , layer (electronics) , linguistics , philosophy , chemistry , organic chemistry
A distinctive Deep Convolution Neural Network (CNN) outline architecture model is presented in this work which proficiently signify complex image features. With two convolution layers implementation of feature extraction is carried out successfully and provides a well-built biometric verification system by using physiological traits Palm print and Hand vein for development of our system. The formation of CNN hyper parameters carried out in numerous experiments such as filters and its size in each layer, number of convolution layers needed, batch size, iterations, learning rate is a paramount and epochs defining these factors are truly dependent on the nature of data and its size. The new method presented in our model has been part with and it gives 99% of GAR in biometric verification of unimodal system and extremely the experimentation has condensed results when combined with deep feature extraction and classification outcomes.