Methods for ultrasound imaging systems to improve image qualities and to reduce the system cost

超声成像是一种安全、相对简单和廉价的用于软组织成像的实时模式，并且已经在临床上广泛使用多年。它具有良好的空间分辨率和成像深度。 然而，低对比度病变的可检测性差和3-D超声成像系统的高成本是超声应用的两个显著限制。申请人的研究计划将解决这两个问题。所提出的工作有两个目标：（1）通过开发先进的图像重建算法来提高病变的可检测性，该算法将处理合成孔径成像（STA）中的预波束形成的射频（RF）通道数据。这将实现更好的图像质量，即空间分辨率和对比噪声比，并且将有助于识别超声图像中的小目标，例如神经、小血管和小肿瘤;（2）通过减少同时接收通道的数量和简化传输系统来降低超声系统成本。申请人的团队在过去三年中取得了一些初步成果，朝着上述两个目标迈进。 在初步的调查中，我们已经证明了所提出的方法的可行性。然而，我们只测试了新的方法在数字幻影和明胶幻影在初步的实验。此外，新方法中的解码算法由于计算复杂度高而对于任何工业或临床应用来说太耗时。 这项拟议的研究是上述项目的自然进展。研究内容包括：（a）加速所提出的算法并探索进一步的方法来提高病变检测能力;（B）考虑像差和频率相关衰减对图像重建算法性能的影响，并在组织模型上测试所提出的方法;（c）将压缩感知（CS）和小波变换应用于合成孔径超声成像，以提高成像系统的性能;（d）将减少接收通道的方法扩展到三维超声成像的模拟。从长远来看，这些输出将推动算法走向未来的临床研究和技术转化。

Ultrasound imaging is a safe, relatively simple and inexpensive, and real-time modality for the imaging of soft tissues, and has been used ubiquitously in clinics for many years. It has good spatial resolution and imaging depth. However, poor detectability of low-contrast lesions and the high cost of 3-D ultrasound imaging systems are two significant limitations of ultrasound applications. The applicant's research program will address both of these. The proposed work has two goals: (1) to improve lesion detectability by developing advanced image reconstruction algorithms that will process the pre-beamformed radiofrequency (RF) channel data in Synthetic Aperture imaging (STA). This will achieve better image quality, i.e. spatial resolutions and contrast-noise-ratio, and will help identify small targets, such as nerve, small blood vessels, and small tumors in ultrasound images; (2) to reduce the ultrasound system cost by lowering the number of simultaneous receiving channels and simplifying the transmission system. The applicant's team has obtained some preliminary results that have advanced toward the above two goals in the last three years. In the preliminary investigations we have demonstrated the feasibility of the proposed methods. However, we have only tested the new methods on numerical phantoms and gelatin phantoms in the preliminary experiments. In addition, the decoding algorithms in the new methods are too time-consuming for any industrial or clinical applications due to the high computational complexity. This proposed research represents a natural progression of the above projects. It includes (a) accelerating the proposed algorithms and exploring further methods to improve lesion detectability; (b) considering the effect of aberration and frequency-dependent attenuation on the performance of the image reconstruction algorithms and testing the methods on tissue phantoms; (c) applying compressive sensing (CS) and wavelet transform to synthetic aperture ultrasound imaging to improve imaging system performance; (d) extending the method of reducing receiving channels to the simulations of 3D ultrasound imaging. In the long term, these outputs will advance the algorithms toward future clinical investigations and technology translation.