IEEE Trans Ultrason Ferroelectr Freq Control. 2021 May;68(5):1628-1640. doi: 10.1109/TUFFC.2020.3043326. Epub 2021 Apr 26.
A 3-D synthetic transmit aperture ultrasound imaging system with a fully addressed array usually leads to high hardware complexity and cost since each element in the array is individually controlled. To reduce the hardware complexity, we had presented the large-pitch synthetic transmit aperture (LPSTA) ultrasound imaging for 2-D imaging using a 1-D phased array to reduce the number of measurement channels M (the product of number of transmissions, [Formula: see text], and the number of receiving channels in each transmission, [Formula: see text]). In this article, we extend this method to a 2-D matrix array for 3-D imaging. We present both numerical simulations and experimental measurements. We combined L × L adjacent elements into transmission subapertures (SAP) and K × K adjacent elements into receive SAPs in synthetic transmit aperture (STA) imaging. In the image reconstruction, we conducted the first attempt to apply and integrate Gaussian-approximated spatial response function (G-SRF) with delay and sum (DAS) to improve the image contrast, especially for the near-field targets. The imaging performance obtained from G-SRF was also evaluated numerically and compared with the previously presented frequency-domain SRF (Freq-domain SRF). The 3-D large-pitch synthetic transmit aperture (3-D-LPSTA) with G-SRF can provide a computationally efficient solution compared with the standard 3-D-STA method. With approximately 1900-fold reduction in the number of measurement channels, 3-D-LPSTA can provide image contrast comparable with the standard 3-D-STA with a full array and significantly better than using a periodically sparse array with similar complexity. In addition to reducing the system complexity, the 3-D-LPSTA achieves 700-fold reduction in computational complexity and 523-fold reduction in data storage. Finally, we evaluated and implemented the G-SRF using phantom data, which were consistent with the simulation results showing that the G-SRF can improve the image contrast. The results demonstrate that the proposed 3-D-LPSTA shows the great potential for designing an inexpensive ultrasound system to ensure the real-time 3-D clinical ultrasound imaging using large arrays. The limits of the proposed method were also discussed.
一种具有完全寻址阵列的三维合成发射孔径超声成像系统通常会导致硬件复杂度和成本增加，因为阵列中的每个元件都需要单独控制。为了降低硬件复杂度，我们提出了使用一维相控阵进行二维成像的大间距合成发射孔径（LPSTA）超声成像，以减少测量通道数 M（发射次数的乘积，[公式：见文本]，以及每次发射中接收通道的数量，[公式：见文本]）。在本文中，我们将该方法扩展到用于三维成像的二维矩阵阵列。我们同时进行了数值模拟和实验测量。我们将 L×L 个相邻元件组合成发射子孔径（SAP），并将 K×K 个相邻元件组合成合成发射孔径（STA）成像中的接收子孔径（SAP）。在图像重建中，我们首次尝试应用和集成高斯逼近空间响应函数（G-SRF）与延迟求和（DAS）以提高图像对比度，特别是对于近场目标。还从数值上评估了 G-SRF 的成像性能，并与之前提出的频域 SRF（频域 SRF）进行了比较。与标准的 3-D-STA 方法相比，具有 G-SRF 的 3-D 大间距合成发射孔径（3-D-LPSTA）可以提供更有效的计算解决方案。通过将测量通道数减少约 1900 倍，3-D-LPSTA 可以提供与使用全阵列的标准 3-D-STA 相当的图像对比度，并显著优于使用类似复杂度的周期性稀疏阵列。除了降低系统复杂度外，3-D-LPSTA 还实现了计算复杂度降低 700 倍和数据存储减少 523 倍。最后，我们使用 phantom 数据评估和实现了 G-SRF，结果与模拟结果一致，表明 G-SRF 可以提高图像对比度。结果表明，所提出的 3-D-LPSTA 为设计廉价的超声系统以确保使用大型阵列进行实时 3-D 临床超声成像提供了巨大的潜力。还讨论了所提出方法的局限性。
