Research And Implementation Of Common Abnormal ECG Signal Classification Methods For Mobile Platforms

Heart disease has received much attention in recent years.Preventing and diagnosing the heart disease as soon as possible can greatly reduce mortality.The ECG signal can characterize the heart activity,and heart health status can be grasped by the ECG signal analysis.At present,there are two problems in the analysis of ECG signals.Firstly,the traditional classification method of ECG signals has the problems of inaccurate waveform positioning and artificial selection of features.Therefore,the correctness of classification cannot be guaranteed.Secondly,it is difficult to migrate to application from the traditional classification algorithms during engineering,and with low applicability.There are few ECG signal analysis systems for mobile platforms at present,and lacking of analysis of abnormal ECG signals.Therefore,developing a mobile platform system for automatic analysis of common abnormal ECG signals,realtime monitoring and analysis of ECG signals has become an urgent problem to be solved.In view of the above problems,comparing a variety of traditional classification algorithms,this thesis uses the deep convolutional neural network algorithm as the basic framework,and improves the algorithm for the data imbalance problem of abnormal ECG signals.An automatic classification algorithm for ECG signals based on weight and deep convolutional neural network.By preprocessing the ECG data of the MITBIH arrhythmia database,the experimental data set training algorithm model is obtained,and the algorithm model is transferred to the project.In the system design and implementation,comparing the advantages and disadvantages of ECG monitoring system platform,according to the realization goal of the system,the overall design of the system was carried out from the functional requirements and performance requirements,and the mobile platform system,realizing the classification of common abnormal ECG signals,is developed.The system is capable of long-term monitoring,and can be used to analyze ECG signals by calling the ECG automatic classification algorithm model.Finally,the implementation of each main function module in the system is introduced in detail,including Bluetooth forwarding module,data receiving module,ECG display module,ECG signal automatic analysis module,etc.Finally,the system development is completed on the Android system platform.In order to verify the rationality and correctness of the logic design of the mobile platform analysis system,this thesis conducts functional logic test and performance test on the main functional modules of the system.The test results show that the system can run stably,and can carry out long-term monitoring of ECG signals.It can correctly call the ECG signal automatic classification algorithm based on deep Convolutional Neural Network,analyze ECG signals,display ECG signal waveforms,analyze of heart beat type,and calculate the average heart rate,meeting the goal of real-time continuous monitoring and analysis of ECG signals.