Increasing Security of Telemedicine Service

Abstract

The paper is about the improving the safety of telemedicine services that carry out the diagnosis of cardiovascular diseases on ECG. There are 2 deep learning methods (Multilayer Perceptron and Recurrent Neural Network), 14 machine learning methods (Naive Bayes classifier for multivariate Bernoulli models, A decision tree classifier, An extremely randomized tree classifier, Classifier implementing the k nearest neighbors vote, Linear Discriminant Analysis, Linear Support Vector Classification, Logistic Regression, Nearest centroid classifier, A random forest classifier, Classifier using Ridge regression, Ridge classifier with built in cross validation, Gaussian Mixture Models, Support Vector Machines), 4 ECG digitization time intervals (5, 10, 15 and 20 seconds), and 3 databases of digitized electrocardiograms (The Physikalisch-Technische Bundesanstalt (PTB) Diagnostic ECG Database, European ST-T Database, St.-Petersburg Institute of Cardiological Technics 12-lead Arrhythmia Database). It was realized that The accuracy of biometric identification and diagnosis of cardiovascular diseases increases with an increase in ECG registration time to about 10 seconds, after which it reaches a plateau, biometric identification and diagnosis of cardiovascular diseases are possible with a signal registration time of 5 seconds and the most stable recognition results were given by such methods of classification of biometric features as fully connected neural network (MLP), An extremely randomized tree classifier and Classifier implementing the knearest neighbors vote.

Biometric Authentication Based on Electrocardiogram

“…We have shown that wavelets such as Haar, Daubechies, and Symlets are best suited for biometric identification. We have shown that good results can be obtained if the entire cardiocycle is used as biometric features [18]. The number of features in this case depends on the sampling frequency of the signal.…”