Research On Dynamic ECG Denoising Algorithm Based On Depth Factor Analysis

With the aging of population and the acceleration of urbanization,more and more people are suffering from heart disease in recent years.Because that static Electrocardiogram(ECG)examination cannot provide enough information for the diagnosis for patients with rational symptom,doctors will recommend dynamic ECG monitoring.Removing the noise from the dynamic ECG and improving the accuracy of automatic detection are gradually becoming a hot research topic in tele-medicine.Factor analysis,which provides a new pointcut for ECG denoising,has been widely used in speech recognition and face recognition.In order to remove the complex noise in ECG,a novel algorithm based on depth factor analysis is proposed in this paper.The main contents are as follows:(1)A model of ECG denoising is developed based on factor analysis.The factor analysis model was adopted into the field of the ECG denoising.Based on the factor analysis model and the ECG database,a machinge learning procedure is developed to obtain the hidden factors of ECG,and remove the noise.(2)An depth factor analysis based ECG denoising algorithm is proposed.Considering that the hidden factor of the ECG is interfered by the noise,we use the method of deepening the hidden factor layer by layer to build the factor analysis.The hidden factors of each layer as input to obtain deep hidden factors and discard the Gaussian noise of each layer.Finally,reconstruct the clean ECG data from the top layer hidden factor.The experimental results on the MIT database show the good performance of the proposed algorithm.(3)The depth factor analysis based ECG denoising algorithm is applied to ECG monitoring platform.In order to verify the research results of this paper,the depth factor analysis denoising algorithm is applied to the intelligent ECG monitoring platform developed by the our team.Based on the data extracted from the platform,the proposed algorithm can be used to filter the complex noise while maintaining the main characteristics of ECG waveform.