Distributed processor for noninvasive blood flow analysis

High-resolution ultrasound imaging of peripheral arteries offers a powerful technique for screening of potential stroke victims. However, the technique becomes inaccurate in cases of low-grade stenosis and the presence of noncalcified plaque. In these cases Doppler flow measurement can be used to augment the imaging system to arrive at a more accurate diagnosis. This paper describes a duplex system that combines high-resolution imaging with Doppler measurements. The various real-time processing and control tasks are implemented in a distributed processing system using three different processors. A high-speed digital signal processor (DSP) performs Fourier transforms on the Doppler-shifted signals and also performs some of the premultiplications required for Doppler parameter computations. A dedicated Doppler CPU receives the spectral coefficients from the DSP and computes several parameters such as gray-scale spectrogram coefficients, frequency centroids, spectral broadening coefficients, and velocity profiles. A third CPU is used to control the imaging system and to perform the housekeeping tasks such as control of the video display and the interface with the control panel.