Research on Personal Identity Verification Based on Convolutional Neural Network

Abstract

In this paper, we propose a Personal Identity Verification (PIV) method based on 2-D convolutional neural network (CNN) by using electrocardiosignal (ECG singles). CNN shows outstanding performance in the field of image recognition nowadays, in order to make better use of this advantage, we innovatively convert electrocardiosignal into 2-D grayscale instead of traditional ECG. While ensuring that the image contains a complete cardiac cycle, it also enables the network to fully learn both the characteristics of the electrocardiosignal period and characteristics between each electrocardiosignal period. Optimization of the proposed CNN classifier includes various deep learning techniques such as batch normalization, data augmentation, Xavier initialization, and dropout. As a result, our classifier achieved 99.90% average accuracy. To precisely validate our CNN classifier, 10-fold cross-validation was performed at the evaluation which involves every ECG recording as a test data. Our experimental results have successfully validated that the proposed CNN classifier with the transformed ECG images can achieve excellent identification accuracy.