Real-time synthetic aperture imaging: opportunities and challenges

Synthetic aperture (SA) ultrasound imaging has not been introduced in commercial scanners mainly due to the computational cost associated with the hardware implementation of this imaging modality. SA imaging redefines the term beamformed line. Since the acquired information comes from all points in the region of interest it is possible to beamform the signals along a desired path, thus, improving the estimation of blood flow. The transmission of coded excitations makes it possible to achieve higher contrast and larger penetration depth compared to "conventional" scanners. This paper presents the development and implementation of the signal processing stages employed in SA imaging: compression of received data acquired using codes, and beamforming. The goal was to implement the system using commercially available field programmable gate arrays. The compression filter operates on frequency modulated pulses with duration of up to 50 mus sampled at 70 MHz. The beamformer can process data from 256 channels at a pulse repetition frequency of 5000 Hz and produces 192 lines of 1024 complex samples in real time. The lines are described by their origin, direction, length and distance between two samples in 3D. This parametric description makes it possible to quickly change the image geometry during scanning, thus enabling adaptive imaging and precise flow estimation. The paper addresses problems such as large bandwidth and computational load and gives the solutions that have been adopted for the implementation.