Extraction And Classification Of Wearable ECG Signal Feature

The incidence of cardiovascular diseases has been increasing in recent years,and cardiovascular disease has become the leading cause of mortality in the world.For this reason,daily ECG monitoring is becoming a research hotspot,and wearable devices play an important role in daily ECG monitoring.Wearable devices mainly use the information of QRS complex and heart rate variability in the ECG signal to assess heart function.Therefore,in this paper,based on the characteristics of ECG signal and application scenarios of wearable devices,QRS complex detection of ECG signal is performed firstly,and then five different classifiers are designed to classify different types of heart rhythm,which are based on time domain characteristics and frequency domain characteristics of heart rate variability.Finally,an ECG monitoring application is designed based on the Android platform to cooperate with the self-developed wearable ECG device.In the application scenario of wearable devices,the user's posture or the noise of the surrounding environment will cause great interference to the detection of the QRS complex.To detect R peak more accurately in wearable devices,especially in exercise,an improved method called ISC algorithm is proposed with high anti-interference ability for R peak detection in wearable devices based on a simple basic algorithm called SC algorithm.The proposed method is characterized by using the updated amplitude selection threshold,updated slope comparison threshold,RR interval judgement and maximum correction to reduce false positives and false negatives,which improves the accuracy of R peak detection and the anti-interference ability of the algorithm effectively.For data from MIT-BIH Arrhythmia Database,the positive predictivity P+ of ISC algorithm can reach 99.12%,and the sensitivity Se of ISC algorithm is more than 95%.For MIT-BIH Noise Stress Test Database,the accuracy of ISC algorithm for both sensitivity Se and positive predictivity P+ can exceed 94%under three common noise,baseline wander,muscle artifact,and electrode motion artifact.For wearable devices in exercise,even under the exercise intensity of 7 km per hour,the average positive predictivity P+ of ISC algorithm is 99.32%.On the basis of R wave detection,Q wave and S wave are detected by differential and maximum value method and maximum value method.The average Se of Q wave detection is 96.02%,and the average P+ is 99.22%.In the meanwhile,the average Se of S wave detection is 95.94%,and the average P+ is 99.17%.The accuracy of Q wave and S wave are equivalent to the R wave detection to detect the position of the QRS complex correctly,which is suitable for real-time QRS complex detection of wearable devices.Heart rate variability can characterize the body's autonomic nervous system as well as cardiac function,generally measured by RR intervals.In this paper,10 features from the time domain features and frequency domain features of five different types of heart rhythm signals are extracted,including the RR interval mean(RR),RR interval standard deviation(SDNN),the rms value of the adjacent cardiac interval are extracted(RNSSD),the standard deviation of the difference between adjacent cardiac intervals(SDSD),the number of cardiac cycles that the adjacent RR intervals is more than 50 ms(NN50),the percentage of NN50 in all cardiac cycles(PNN50),TP,LF,HF and LF/HF.Then five heart rhythm classifiers of two,three,four and five types classifications are constructed respectively based on BP neural network.Data from MIT-BIH normal sinus rhythm database,BIDMC congestive heart failure database and MIT-BIH arrhythmia in PhysioBank were used to train and test.The average accuracy of the five-class and four-class heart rhythm classifiers based on BP neural network is 68.514%and 73.790%,respectively.The average accuracy of the two-class heart rhythm classifier to distinguish normal sinus rhythm and congestive heart failure is 89.448%,which can provide effective warning for heart failure patient and has high application value.Finally,based on the Android platform,an ECG monitoring application is designed,integrating many functions such as realtime processing of ECG data,real-time display of ECG waveform,call up emergency contact,data saving and exporting of ECG,and cooperated with self-developed wearable ECG device,which is flexible and convenient to use.