An Approach for Biometric Verification Based on Human Body Communication using Convolutional Neural Network

In this paper, a new approach based on human body communication (HBC) was presented for biometric verification. Specifically, the transmission gain S21 of human forearm is regarded as the biometric trait. For this purpose, three different forearm models were established and proposed to validate the aforementioned approach. Furthermore, the transmission gain S21 of 21 volunteers was obtained by the use of vector network analyzer (VNA) in the frequency range 0.3 MHz to 1500 MHz in the experiment. In addition, the convolutional neural network (CNN) which includes 3 convolution layers, 3 max pooling layers and 2 fully-connected layers was adopted in this paper. The influences of different optimizers and loss functions on CNN were investigated. The results showed that the recognition accuracy of CNN was 99.86% when the optimizer was set as Adadelta and the loss function was set as categorical crossentropy. We therefore suggest that the CNN has potentials to improve the recognition accuracy of biometric verification based on HBC.