Coded excitation and sub-band processing for blood velocity estmation in medical ultrasound

This paper investigates the use of broadband coded excitation and subband processing for blood velocity estimation in medical ultrasound. In conventional blood velocity estimation a long (narrow-band) pulse is emitted and the blood velocity is estimated using an auto-correlation based approach. However, the axial resolution of the narrow-band pulse is too poor for brightness-mode (B-mode) imaging.

Therefore, a separate transmission sequence is used for updating the B-mode image, which lowers the overall frame-rate of the system. By using broad-band excitation signals, the backscattered received signal can be divided into a number of narrow frequency bands. The blood velocity can be estimated in each of the bands and the velocity estimates can be averaged to form an improved estimate.

Furthermore, since the excitation signal is broadband, no secondary B-mode sequence is required, and the frame rate can be increased. To increase the SNR for the broad-band excitation waveforms, coding is proposed. Three different coding methods are investigated: nonlinear frequency modulation, complementary (Golay) codes, and Barker codes.

Code design is described for the three different methods as well as different ways of pulse compression for restoring axial resolution. The different methods were studied using an experimental ultrasound scanner and a circulating flow rig.