COMPUTING REQUIREMENTS OF MODERN MEDICAL DIAGNOSTIC ULTRASOUND MACHINES

Ultrasound is one of the leading medical imaging modalities because it is safe, noninvasive, portable, easy to use, relatively inexpensive a nd displays images in real-time. Due to its real-time nature, an ultra sound machine must be able to process its incoming data quickly. High computational and throughput requirements in modern ultrasound machine s have restricted their internal design to algorithm-specific hardware with limited programmability. Adding new ultrasound imaging applicati ons or improving a machine's internal algorithms can require costly ha rdware redesigns and replacements of boards or of the entire machine. In an effort to address these problems, we have reviewed each of the e ssential functions in modern ultrasound machines and analyzed their co mputational requirements in programmable systems. These functions incl ude dynamic downconversion, tissue signal processing, color flow proce ssing, scan conversion, and tissue/flow decision. Our estimate of the total computing requirement to current ultrasound machines is 54.71 bi llion operations per second (BOPS) when transcendental functions are i mplemented in software and 31.26 BOPS when transcendental functions ar e implemented in lookup tables taking 160.26 Mbytes of memory. To inve stigate the feasibility of programmable generalized ultrasound systems , we have designed a flexible and parallel programmable ultrasound pro cessing subsystem, called the Programmable Ultrasound Image Processor (PUIP). As the need for programmable ultrasound machines increases in the future due to various advantages (e.g., lower system cost, faster clinical use and lower research/development expenses), it will be cruc ial to develop not only a high-performance, scalable, reconfigurable p arallel computer architecture meeting the computing requirement, but a lso efficient ultrasound algorithms that can be optimally mapped into the parallel architecture. (C) 1998 Published by Elsevier Science B.V. All rights reserved.