Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing 100084, China.
Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing 100084, China; Tsinghua-Peking Joint Center for Life Sciences Department, Tsinghua University, Beijing 100084, China.
Med Image Anal. 2021 May;70:102018. doi: 10.1016/j.media.2021.102018. Epub 2021 Feb 25.
Image reconstruction from radio-frequency (RF) data is crucial for ultrafast plane wave ultrasound (PWUS) imaging. Compared with the traditional delay-and-sum (DAS) method based on relatively imprecise assumptions, sparse regularization (SR) method directly solves the inverse problem of image reconstruction and has presented significant improvement in the image quality when the frame rate remains high. However, the computational complexity of SR is too high for practical implementation, which is inherently associated with its iterative process. In this work, a deep neural network (DNN), which is trained with an incorporated loss function including sparse regularization terms, is proposed to reconstruct PWUS images from RF data with significantly reduced computational time. It is remarkable that, a self-supervised learning scheme, in which the RF data are utilized as both the inputs and the labels during the training process, is employed to overcome the lack of the "ideal" ultrasound images as the labels for DNN. In addition, it has been also verified that the trained network can be used on the RF data obtained with steered plane waves (PWs), and thus the image quality can be further improved with coherent compounding. Using simulation data, the proposed method has significantly shorter reconstruction time (∼10 ms) than the conventional SR method (∼1-5 mins), with comparable spatial resolution and 1.5-dB higher contrast-to-noise ratio (CNR). Besides, the proposed method with single PW can achieve higher CNR than DAS with 75 PWs in reconstruction of in-vivo images of human carotid arteries.
从射频 (RF) 数据进行图像重建对于超快速平面波超声 (PWUS) 成像至关重要。与基于相对不精确假设的传统延迟求和 (DAS) 方法相比，稀疏正则化 (SR) 方法直接求解图像重建的逆问题，在保持高帧率时，图像质量有了显著提高。然而，SR 的计算复杂度对于实际实现来说太高，这与其迭代过程有关。在这项工作中，提出了一种深度神经网络 (DNN)，它使用包含稀疏正则化项的合并损失函数进行训练，可从 RF 数据中重建 PWUS 图像，同时显著减少计算时间。值得注意的是，采用了一种自监督学习方案，在训练过程中，RF 数据既用作输入，也用作标签，以克服缺乏“理想”超声图像作为 DNN 标签的问题。此外，还验证了所训练的网络可用于定向平面波 (PWs) 获得的 RF 数据，从而可以通过相干复合进一步提高图像质量。使用仿真数据，与传统的 SR 方法 (∼1-5 分钟) 相比，所提出的方法的重建时间显著缩短 (∼10 毫秒)，具有可比的空间分辨率和 1.5 分贝更高的对比度噪声比 (CNR)。此外，与使用 75 个 PW 的 DAS 相比，在重建人体颈动脉的体内图像时，单 PW 的提出的方法可获得更高的 CNR。
