Estimation Of Ultrasonic Doppler Blood Flow Velocity Distribution Based On EWT Time-Freauency Analysis

The velocity distribution of arterial blood flow is one of the important parameters for evaluating human physiology and pathology.It not only reflects the body’s metabolic level but also contributes to the diagnosis of cardiovascular diseases.Therefore,the measurement of blood flow velocity distribution plays a significant role in clinical diagnosis.Ultrasound Doppler shift method is widely used in clinical blood flow velocity detection because of its non-invasive,real-time;etc.The method obtains signals involving a Doppler frequency shift information of blood flow signal by ultrasonic scattering of red blood cells in the blood.This signal is a non-stationary random signal subject to Gaussian distribution，and its time-frequency distribution is closely related to blood flow velocity.Short-time Fourier transformation(STFT)is the common method for estimating the time-frequency distribution of ultrasound Doppler blood flow signals.However,this method requires that the short-time stability of the signal.In addition,this method can not take into account both time and frequency resolution.As a result,the traditional time-frequency analysis method based on STFT has a problem of low accuracy in estimating the blood flow velocity distribution.In order to quantitatively analyze the ultrasound Doppler blood flow signal more accurately for the diagnosis of cardiovascular disease,this paper proposes a time-frequency analysis method based on the empirical wavelet transform(EWT)algorithm.This method firstly adaptively divides the Fourier spectrum of the signal and then establishes a suitable wavelet filter group to extract the target single-component components.Finally,the Hilbert transform is performed on each single-component component to obtain the instantaneous frequency and the instantaneous amplitude.As a result,this method can get a better time-frequency resolution.In order to verify the feasibility of the EWT-based blood flow velocity distribution estimation method,this paper will be based on the Field Ⅱ carotid ultrasound RF blood flow signal simulation model.Firstly,select the blood blow echo signals at different radial positions in the blood vessel.Secondly,use the traditional STFT and the EWT-based time-frequency analysis method estimates to the signals to estimate blood flow velocity distribution respectively.Finally,calculate their normalized root mean square errors(NRMSEs)to be 0.18 and 0.12,Afterward,analysis and comparison show that the accuracy of the proposed method is improved by 33%.In summary,the EWT-based time-frequency analysis method proposed in this paper can provide a better time-frequency resolution,and it makes extraction of Doppler shift information more precise.Therefore,it makes the blood flow velocity distribution estimation more accurate.