Ali Rehman, Mitcham Trevor, Brickson Leandra, Hu Wentao, Doyley Marvin, Rubens Deborah, Ignjatovic Zeljko, Duric Nebojsa, Dahl Jeremy
University of Rochester Medical Center, Department of Imaging Sciences, Rochester, New York, United States.
Stanford University School of Medicine, Department of Radiology, Palo Alto, California, United States.
J Med Imaging (Bellingham). 2022 Nov;9(6):067001. doi: 10.1117/1.JMI.9.6.067001. Epub 2022 Nov 1.
Isolating the mainlobe and sidelobe contribution to the ultrasound image can improve imaging contrast by removing off-axis clutter. Previous work achieves this separation of mainlobe and sidelobe contributions based on the covariance of received signals. However, the formation of a covariance matrix at each imaging point can be computationally burdensome and memory intensive for real-time applications. Our work demonstrates that the mainlobe and sidelobe contributions to the ultrasound image can be isolated based on the receive aperture spectrum, greatly reducing computational and memory requirements.
The separation of mainlobe and sidelobe contributions to the ultrasound image is shown in simulation, , and using the aperture spectrum method and multicovariate imaging of subresolution targets (MIST). Contrast, contrast-to-noise-ratio (CNR), and speckle signal-to-noise-ratio are used to compare the aperture spectrum approach with MIST and conventional delay-and-sum (DAS) beamforming.
The aperture spectrum approach improves contrast by 1.9 to 6.4 dB beyond MIST and 8.9 to 13.5 dB beyond conventional DAS B-mode imaging. However, the aperture spectrum approach yields speckle texture similar to DAS. As a result, the aperture spectrum-based approach has less CNR than MIST but greater CNR than conventional DAS. The CPU implementation of the aperture spectrum-based approach is shown to reduce computation time by a factor of 9 and memory consumption by a factor of 128 for a 128-element transducer.
The mainlobe contribution to the ultrasound image can be isolated based on the receive aperture spectrum, which greatly reduces the computational cost and memory requirement of this approach as compared with MIST.
分离主瓣和旁瓣对超声图像的贡献，可通过去除离轴杂波来提高成像对比度。先前的工作基于接收信号的协方差实现了主瓣和旁瓣贡献的这种分离。然而，对于实时应用而言，在每个成像点形成协方差矩阵可能在计算上非常繁重且内存需求很大。我们的工作表明，基于接收孔径谱可以分离主瓣和旁瓣对超声图像的贡献，从而大大降低计算和内存需求。
使用孔径谱方法和亚分辨率目标多协变量成像（MIST），在模拟中展示了主瓣和旁瓣对超声图像贡献的分离。使用对比度、对比噪声比（CNR）和散斑信噪比来比较孔径谱方法与MIST以及传统延迟求和（DAS）波束形成。
孔径谱方法比MIST提高对比度1.9至6.4dB，比传统DAS B模式成像提高8.9至13.5dB。然而，孔径谱方法产生的散斑纹理与DAS相似。因此，基于孔径谱的方法的CNR比MIST小，但比传统DAS大。对于128阵元换能器，基于孔径谱方法的CPU实现显示计算时间减少了9倍，内存消耗减少了128倍。
基于接收孔径谱可以分离主瓣对超声图像的贡献，与MIST相比，这大大降低了该方法的计算成本和内存需求。
