Apodizing Delay and Auto-Correlation Reconstruction Algorithm for High-Frame-Rate Ultrasound Imaging

Abstract

Delay and sum (DAS) beamforming is a common technique for ultrasound imaging, as it has a low computation complexity suitable for real-time imaging. However, the image quality of DAS is largely restricted by the imaging array configuration, such as the aperture size and the center frequency used. Adaptive beamforming techniques such as minimal variance (MV), which presents improved spatial resolution and contrast ratio, but with an significant increased computation requirements that can not generally be implemented with real-time monitoring. In this study, we developed an alternative nonlinear beamforming method for ultrasound imaging, termed an apodizing delay and auto-correlation reconstruction algorithm (ADAC). The computation complexity of ADAC is O(2N) which ensures a similar beamforming speed to DAS O(N), since both are on the same order of magnitude. We tested ADAC against DAS, on both Field II simulated data and experimental data acquired from imaging a standard CIRS calibration phantom. Results show that the ADAC beamformer compared with DAS that mitigated the self-nulling phenomenon of CIRS proximal targets and showed the weak-signal mapping ability up to an imaging depth of 65 mm. The image quality beamformed by ADAC from anechoic simulated data and interleaving B-mode showed an improved contrast ratio (CR), lateral, and axial resolution of 19 dB, 0.5 mm (lateral) and 0.3 mm mm (axial), respectively. All above demonstrate that the ADAC beamformer outperforms DAS with narrower main lobes and increased dynamic range whilst maintaining a similar computational complexity.