Adaptive signal processing in medical ultrasound beamforming

TLDR: In this article, the performance of four beamforming algorithms (Frost BF, Duvall BF, SSB, and SPOC) was compared to the conventional, data independent, beamforming.

Abstract: For over thirty years adaptive beamforming (AB) algorithms have been applied in RADAR and SONAR signal processing. Higher resolution and contrast is attainable using those algorithms at the price of an increased computational load. In this paper we consider four beamformers (BFs): Frost BF, Duvall BF, SSB, and SPOC. These algorithms are well know in the RADAR/SONAR literature. We have performed a series of simulations using ultrasound data to test the performance of those algorithms and compare them to the conventional, data independent, beamforming. Every algorithm was applied on single channel ultrasonic data that was generated using Field II. For a 32 element linear array operating at 5 MHz, beamplot results show that while the Duvall and SSB beamformers reduce sidelobes by roughly 20 dB, the sidelobes using the Frost algorithm rise by 23dB. The -6dB resolution is improved by 38%, 83%, and 43% in the case of Duvall, Frost, and SSB algorithms, respectively. In the case of SPOC, the beamplot shows a super-resolution peak with noise floor at -110 dB. Similar results were obtained for an array consisting of 64 elements.