Research On ECG Noise Reduction Based On The Diffusion Model

Electrocardiogram?ECG?is a commonly used non-invasive detection method in the diagnosis of cardiovascular diseases.The doctor can diagnose the heart condition by watching the electrocardiogram signal recorded by the electrocardiograph machine.Because the ECG signal is a weak physiological signal with low amplitude,low frequency,and low signal-to-noise ratio,the weak signal is easily polluted by various noises during the acquisition process.Therefore,research on ECG signal noise reduction technology has important research significance for improving the reliability of ECG diagnosis.The main work of this paper is as follows:?1?The diffusion function of the nonlinear diffusion model can carry the feedback of peak position,it has a better ability to maintain peaks.Therefore,the research on ECG noise reduction based on classical nonlinear diffusion model has been carried out.First,the explicit difference scheme and semi-implicit difference scheme of the classical nonlinear diffusion model are given.In order to avoid the inversion matrix operation,the"catch-up method"is used to rewrite the semi-implicit difference format to break the limitation of the algorithm on the signal length.Secondly,the design of the diffusion function,three types of diffusion functions are given and the corresponding effectiveness are analyzed.A method for setting global parameters is given through the discussion about the parameter setting of?.Then,a method for setting the diffusion function in sections is given by QRS wave detection.Finally,the effectiveness of the algorithm is verified and compared with the existing main methods,the results show that the classical nonlinear diffusion model has better results for most signals than the other methods about the two indicators of improved signal-to-noise ratio(SNR_{im p})and mean square error?MSE?.?2?A large number of literatures show that the fractional model has better performance than the integer model.Therefore,the research on ECG noise reduction based on time fractional nonlinear diffusion model has been carried out.The time fractional nonlinear diffusion model is obtained by replacing the time derivative on the left side of the classical nonlinear diffusion model with a fractional derivative.When the order of the left time fractional derivative is less than 1,it is called a sub-diffusion model.When the left time fractional derivative is greater than 1,it is called a super-diffusion model.Explicit and semi-implicit algorithms for the time fractional sub-diffusion model and the super-diffusion model are given in this paper,respectively.And the effectiveness of explicit and semi-implicit algorithms for the time fractional sub-diffusion model and the super-diffusion model is verified.The effects of fractional order on noise reduction performance are discussed.Experiments show that when the time fractional order is 1.05 to 1.1,the time fractional nonlinear diffusion model is better than the classic nonlinear diffusion model.Finally,the effectiveness of the algorithm is verified and compared with classical nonlinear diffusion model and the existing main methods,the results show that the time fractional nonlinear diffusion model has better results for all signals than the other methods about the two indicators of improved signal-to-noise ratio(SNR_{im p})and mean square error?MSE?.