IEEE Trans Med Imaging. 2017 Jun;36(6):1372-1381. doi: 10.1109/TMI.2017.2687400. Epub 2017 Mar 24.
The development of ultrafast ultrasound imaging offers great opportunities to improve imaging technologies, such as shear wave elastography and ultrafast Doppler imaging. In ultrafast imaging, there are tradeoffs among image signal-to-noise ratio (SNR), resolution, and post-compounded frame rate. Various approaches have been proposed to solve this tradeoff, such as multiplane wave imaging or the attempts of implementing synthetic transmit aperture imaging. In this paper, we propose an ultrafast synthetic transmit aperture (USTA) imaging technique using Hadamard-encoded virtual sources with overlapping sub-apertures to enhance both image SNR and resolution without sacrificing frame rate. This method includes three steps: 1) create virtual sources using sub-apertures; 2) encode virtual sources using Hadamard matrix; and 3) add short time intervals (a few microseconds) between transmissions of different virtual sources to allow overlapping sub-apertures. The USTA was tested experimentally with a point target, a B-mode phantom, and in vivo human kidney micro-vessel imaging. Compared with standard coherent diverging wave compounding with the same frame rate, improvements on image SNR, lateral resolution (+33%, with B-mode phantom imaging), and contrast ratio (+3.8 dB, with in vivo human kidney micro-vessel imaging) have been achieved. The f-number of virtual sources, the number of virtual sources used, and the number of elements used in each sub-aperture can be flexibly adjusted to enhance resolution and SNR. This allows very flexible optimization of USTA for different applications.
超快超声成象的发展为改善成像技术提供了极好的机会,如剪切波弹性成像和超快多普勒成像。在超快成象中,图像信噪比(SNR)、分辨率和后置复合帧率之间存在折衷。已经提出了各种方法来解决这种折衷,例如多平面波成象或尝试实现合成发射孔径成象。在本文中,我们提出了一种使用具有重叠子孔径的 Hadamard 编码虚拟源的超快合成发射孔径(USTA)成象技术,以在不牺牲帧率的情况下提高图像 SNR 和分辨率。该方法包括三个步骤:1)使用子孔径创建虚拟源;2)使用 Hadamard 矩阵对虚拟源进行编码;3)在不同虚拟源的传输之间添加短时间间隔(几微秒),以允许重叠子孔径。USTA 在点目标、B 模式体模和活体人肾微血管成象中进行了实验测试。与具有相同帧率的标准相干发散波复合相比,图像 SNR、横向分辨率(B 模式体模成象时提高了+33%)和对比比(活体人肾微血管成象时提高了+3.8 dB)都有所提高。虚拟源的 f 数、使用的虚拟源数量以及每个子孔径中使用的元件数量可以灵活调整,以提高分辨率和 SNR。这使得 USTA 可以针对不同的应用进行非常灵活的优化。