Tackling the Bottleneck of Delay Tables in 3D Ultrasound Imaging

Abstract

Abstract-3D ultrasound imaging is quickly becoming a reference technique for high-quality, accurate, expressive diagnostic medical imaging. Unfortunately, its computation requirements are huge and, today, demand expensive, power-hungry, bulky processing resources. A key bottleneck is the receive beamforming operation, which requires the application of many permutations of fine-grained delays among the digitized received echoes. To apply these delays in the digital domain, in principle large tables (billions of coefficients) are needed, and the access bandwidth to these tables can reach multiple TB/s, meaning that their storage both on-chip and off-chip is impractical. However, smarter implementations of the delay generation function, including forgoing the tables altogether, are possible. In this paper we explore efficient strategies to compute the delay function that controls the reconstruction of the image, and present a feasibility analysis for an FPGA platform.

Apodization scheme for hardware-efficient beamformer

“…Since this computation involves the evaluation of complex functions like square roots, it is mandatory to identify accurate, fast and low-area approximation circuits [19], [20]. In particular, we have shown [5], [6] how a highly efficient architecture can be devised by accepting some inaccuracy in the calculation of the delays. The kernel of ultrasound beamforming includes also the step of apodization.…”

“…In our previous papers [5], [6] we have proposed a significant algorithmic simplification that relies on a Taylor expansion of the square root calculation. This method approximates the delay to 1 :…”

“…As the delay values are used to select a sample of the received echoes, this error equates to picking echo signals almost 100 samples off [5], [6], and can essentially be characterized as a noise increase. We have further shown [6] that a way to get rid of this effect is, at those locations S, to ignore the element D echoes that are most impacted by the delay calculation inaccuracy.…”

“…In our previous work [5], [6], we tackled the problem of delay calculation, showing an approach that reduces dramatically the computation cost at the price of a degree of inaccuracy. This inaccuracy can be seen as resulting in added image noise.…”

Single-FPGA, scalable, low-power, and high-quality 3D ultrasound beamformer

“…An approach has been proposed [3] that benefits from the first order Taylor approximation of the square root operation. This approach simplifies the delay calculation to:…”