P6B-9 Clutter Suppression in Doppler Ultrasound Using Wiener Filtering

In Doppler ultrasound as applied to blood flow estimation, pulsatile movement of the vessel wall and surrounding tissue result in a Doppler clutter signal whose spectrum changes over the cardiac cycle and which, when the sample volume (SV) is close to the arterial wall, can cause serious difficulties. Because the clutter signal can be 40 dB greater than that from the red blood cells (RBC) and its time varying spectrum can overlap with the spectrum of the desired signal, the clutter becomes very difficult to identify and remove. This paper proposes a new method for clutter suppression based on the use of information from two sample volumes and Weiner filtering. A Doppler simulation model was developed and used to identify the spectra in two SVs, the primary SV within the vessel lumen and a secondary SV centered within the arterial wall. Data from both SVs were obtained from every backscattered pulse using two range gates. In the primary SV, the clutter signal strength is significantly greater than that from the blood when it is located near the arterial wall However, a secondary SV placed within the arterial wall will contain very little blood signal as the lumen will be within the attenuated portion of the SV energy distribution and the RBC backscattered intensity is already weak. The secondary SV can therefore be assumed to contain just the clutter signal. Wiener filtering was then applied to recover the blood signal from the primary SV. The Wiener filter uses the auto-correlation of the secondary SV and the cross-correlation between the SVs to estimate the clutter signal in the primary SV. This estimated clutter signal was then subtracted from the primary SV to recover the RBC spectrum. Preliminary simulation results using the Wiener filter have shown good agreement to the traditional high-pass FIR filtering method for clutter-to-RBC signal ratios up to 40 dB. Use of the Wiener filtering technique can have significant advantages over traditional filtering methods - due to its adaptability to spectral variation over the cardiac cycle.