ECG Signal Analysis And Software Implementation

Analysis and processing of ECG signal as a multidisciplinary subject is one of the hotspots in signal analysis,which has broad research content,also lots of theory and technology involved.There are still some spaces,by far,for us to improve theoretical research and practical software.The research of ECG signal processing and software implementation mainly includes signal preprocessing,ECG waveform detection and accurate positioning,selection of characteristic parameters and precise extraction and automatic classification of heartbeat type.According to the four aspects of mentioned before,this article are doing some researches as well as discuss the difficulties and key technologies.The main contents of this paper are as follows:(1)Signal preprocessing: In this paper,through the analysis of ECG signal processing and realize of software that is use to fulfill the overall functional required analysis,and insure the pre-processing algorithm design requirements.The preprocessing algorithm requires the removal of most of the high frequency noise,low frequency noise and artifact interference,making the ECG waveforms after pretreatment smooth,continuous,characteristic waveforms.Therefore,this paper uses a real-time noise reduction algorithm based on wavelet transform.(2)ECG waveform detection and accurate positioning: The key to achieve the automatic classification of heartbeat is ECG detection and accurate positioning.According to the requirement of follow-up application of ECG signal,this paper designs QRS complex detection algorithm based on filter bank and adaptive QRS-T based on wavelet transform cancellation P wave detection algorithm,in which P wave detection algorithm and T wave detection algorithm similar.(3)Selection of Characteristic Parameters and Precise Extraction: The efficiency and accuracy of automatic classification of heartbeats depend on a large extent on the selection and extraction of feature parameters.In this paper,the selection and extraction of feature parameters based on deep belief network are used in view of the current research situation.(4)Automatic classification of heartbeats: The automatic classification of the heartbeat is an intuitive representation of the software implementation.According to the characteristic parameters which are come from deep belief network this paper use a class algorithm of nonlinear support vector machine based on Gaussian Kernel function.In terms of the four aspects,this paper designs the module of ECG signal automatic analysis,and other relative software which are needed by the functional modules.Finally,the function of the module is verified by the data of the MIT-BIH database.