Application Of Neural Network Algorithm In Electromagnetic Leakage And Eeg Cognition

Today,with the rapid development of modern medical technology,computer network technology and modern communication technology,telemedicine has become a hot area nowadays.However,as the information security environment becomes more and more complicated,the information transmitted by the telemedicine system will also be leaked,leading to the leakage of personal information and being used by the conscientious people with various problems.Therefore,it is very necessary and important to study the leakage of electromagnetic radiation in telemedicine systems.Studying the principle of electromagnetic radiation leakage can make us better prevent leakage of electromagnetic emission and control the source of electromagnetic radiation leakage.This paper firstly studies the structure of USB data line and the theoretical model of USB data line transmission,and then analyzes and organizes the data we want to transmit according to the transmission protocol to get our original data format.Then the leakage of electromagnetic radiation,which is mainly the USB data line inside the computer,is studied,including the principle of leakage and the way of leakage,which lays the foundation for the comparison between the subsequent leakage data and the original data.Finally,based on the method of machine learning,the classification and recognition of EEG signals in sports imaging are studied.The main innovations of the dissertation are:(1)Designing a USB data electromagnetic leakage acquisition and analysis scheme.(2)Based on the echo state network and support vector machine,the feature extraction and identification of USB electromagnetic leakage is studied.Through the pattern recognition of the collected electromagnetic radiation leakage signal,the highest accuracy can reach 96%in the two classifications.(3)In the analysis of EEG signals in sports imaging,a feature extraction method based on echo state network and independent component analysis is proposed,and the pattern recognition of EEG is studied using support vector machine.The simulation shows that the proposed feature extraction method has significantly improved recognition accuracy compared with the traditional PCA method and ESN feature extraction method.