3D Synthetic Aperture Imaging Method in Spectrum Domain for Low-Cost Portable Ultrasound Systems

Abstract

Portable, hand-held ultrasound devices capable of 3D imaging in real time are the next generation of the medical imaging apparatus adapted not only for professional medical stuff but for a wide group of less advanced users. Limited power supply capacity and the relatively small number of channels used for the ultrasound data acquisition are the most important limitations that should be taken into account when designing such devices and when developing the corresponding image reconstruction algorithms.The aim of this study was to develop a new 3D ultrasound imaging method which would take into account the above-mentioned features of the new generation of ultrasonic devices -low-cost portable general access scanners.It was based on the synthetic transmit aperture (STA) method combined with the Fourier spectrum domain (SD) acoustic data processing. The STA using a limited number of elements in transmit and receive modes for ultrasound data acquisition allowed both aforementioned constraints to be obeyed simultaneously. Moreover, the computational speed of the fast Fourier transform (FFT) algorithm utilized for the ultrasound image synthesis in the spectrum domain makes the proposed method to be more competitive compared to the conventional time domain (TD) STA method based on the delay-and-sum (DAS) technique, especially in the case of 3D imaging in real time mode.Performance of the proposed method was verified using numerical 3D acoustic data simulated in the Field II program for MATLAB and using experimental data from the custom design 3D scattering phantom collected by means of the Verasonics Vantage 256™ research ultrasound system equipped with the dedicated 1024-element 2D matrix transducer.The method proposed in this paper was about 80 times faster than its counterpart based on the time domain synthetic transmit aperture (TD-STA) approach in the numerical example of a single 3D ultrasound image synthesized from 4 partial images each containing 64 × 64 × 512 pixels.It was also shown that the acceleration of the image reconstruction was achieved at the cost of a slight deterioration in the image quality assessed by the contrast and contrast-to-noise ratio (CNR).