Biomedical imaging is unequivocally dependent on the ability to reconstruct interpretable and high-quality images from acquired sensor data. This reconstruction process is pivotal across many applications, spanning from magnetic resonance imaging to ultrasound imaging. While advanced data-adaptive reconstruction methods can recover much higher image quality than traditional approaches, their implementation often poses a high computational burden. In ultrasound imaging, this burden is significant, especially when striving for low-cost systems, and has motivated the development of high-resolution and high-contrast adaptive beamforming methods. Here we show that deep neural networks, that adopt the algorithmic structure and constraints of adaptive signal processing techniques, can efficiently learn to perform fast high-quality ultrasound beamforming using very little training data. We apply our technique to two distinct ultrasound acquisition strategies (plane wave, and synthetic aperture), and demonstrate that high image quality can be maintained when measuring at low data-rates, using undersampled array designs. Beyond biomedical imaging, we expect that the proposed deep learning based adaptive processing framework can benefit a variety of array and signal processing applications, in particular when data-efficiency and robustness are of importance.

Ultrasound localization microscopy has enabled super-resolution vascular imaging through precise localization of individual ultrasound contrast agents (microbubbles) across numerous imaging frames. However, analysis of high-density regions with significant overlaps among the microbubble point spread responses yields high localization errors, constraining the technique to low-concentration conditions. As such, long acquisition times are required to sufficiently cover the vascular bed. In this work, we present a fast and precise method for obtaining super-resolution vascular images from high-density contrast-enhanced ultrasound imaging data. This method, which we term Deep Ultrasound Localization Microscopy (Deep-ULM), exploits modern deep learning strategies and employs a convolutional neural network to perform localization microscopy in dense scenarios, learning the nonlinear image-domain implications of overlapping RF signals originating from such sets of closely spaced microbubbles. Deep-ULM is trained effectively using realistic on-line synthesized data, enabling robust inference in-vivo under a wide variety of imaging conditions. We show that deep learning attains super-resolution with challenging contrast-agent densities, both in-silico as well as in-vivo. Deep-ULM is suitable for real-time applications, resolving about 70 high-resolution patches (128×128 pixels) per second on a standard PC. Exploiting GPU computation, this number increases to 1250 patches per second.

Deep learning for ultrasound image formation is rapidly garnering research support and attention, quickly rising as the latest frontier in ultrasound image formation, with much promise to balance both image quality and display speed. Despite this promise, one challenge with identifying optimal solutions is the absence of unified evaluation methods and datasets that are not specific to a single research group. This article introduces the largest known international database of ultrasound channel data and describes the associated evaluation methods that were initially developed for the challenge on ultrasound beamforming with deep learning (CUBDL), which was offered as a component of the 2020 IEEE International Ultrasonics Symposium. We summarize the challenge results and present qualitative and quantitative assessments using both the initially closed CUBDL evaluation test dataset (which was crowd-sourced from multiple groups around the world) and additional in vivo breast ultrasound data contributed after the challenge was completed. As an example quantitative assessment, single plane wave images from the CUBDL Task 1 dataset produced a mean generalized contrast-to-noise ratio (gCNR) of 0.67 and a mean lateral resolution of 0.42 mm when formed with delay-and-sum beamforming, compared with a mean gCNR as high as 0.81 and a mean lateral resolution as low as 0.32 mm when formed with networks submitted by the challenge winners. We also describe contributed CUBDL data that may be used for training of future networks. The compiled database includes a total of 576 image acquisition sequences. We additionally introduce a neural-network-based global sound speed estimator implementation that was necessary to fairly evaluate the results obtained with this international database. The integration of CUBDL evaluation methods, evaluation code, network weights from the challenge winners, and all datasets described herein are publicly available (visit https://cubdl.jhu.edu for details).

Fetal well-being is commonly assessed by monitoring the fetal heart rate (fHR). In clinical practice, the de-facto standard technology for fHR monitoring is based on Doppler ultrasound (US). Continuous monitoring of the fHR before and during labor is performed using a US transducer fixed on the maternal abdomen. The continuous fHR monitoring, together with simultaneous monitoring of the uterine activity, is referred to as cardiotocography (CTG). In contrast, for intermittent measurements of the fHR, a handheld Doppler US transducer is typically used. In this review paper, the technology of Doppler US for continuous fHR monitoring and intermittent fHR measurements is described, with emphasis on fHR monitoring for CTG. Special attention is dedicated to the measurement environment, which includes the clinical setting in which fHR monitoring is commonly performed. In addition, to understand the signal content of acquired Doppler US signals, the anatomy and physiology of the fetal heart and the surrounding maternal abdomen are described. The challenges encountered in these measurements have led to different technological strategies, which are presented and critically discussed, with focus on US transducer geometry, Doppler signal processing, and fHR extraction methods.

Jects performed five trials of isometric elbow flexion and five trials of elbow extension with increasing levels of force in two conditions: vibration (V) and normal loading (C). V stimulation was characterized by a frequency of 28 Hz. Surface EMG activity of biceps and triceps muscles was simultaneously measured by bipolar surface electromyography and assessed by the estimation of the root mean square (RMS) of the electrical recordings over a fixed 5-s interval. Frequency analysis was adopted to estimate the RMS related to muscle activation and to exclude the harmonics generated by movement artifacts due to V. Results: The analysis of the recordings revealed a significant EMGRMS increase when V was applied. On average, the EMGRMS of biceps and triceps during elbow flexion was, respectively, 26.1% (P < 0.05) and 18.2% (P = 0.15) higher than C. During elbow extension, the EMGRMS of biceps and triceps was 77.2% and 45.2% (P < 0.05) higher than C, respectively. The coactivation was assessed as the ratio between the activation of antagonist and agonist muscles during arm flexion and extension tasks. The results revealed an increase of coactivation during V exercise, especially for lighter loads. Conclusion: This study shows that V exercise at 28 Hz produces an increase of the activation and the coactivation of biceps and triceps. This exercise modality seems therefore suitable for various applications....