Ultrasound Imaging
Overview
Ultrasound imaging studies how to form and interpret images of the body using reflected sound, including B-mode, Doppler, contrast-enhanced, and volumetric scans. Machine-learning approaches range from U-Net-style segmentation and video models to self-supervised and vision-language foundation models, evaluated on tasks such as cardiac function, fetal plane detection, and breast-lesion classification. Physics-informed reconstruction and denoising, alongside robotic scanning and probe guidance, aim to improve image quality and make examinations more consistent and accessible. A central challenge is maintaining reliable performance across devices, hospitals, operators, and patient groups while meeting clinical needs for interpretability, external validation, and efficient deployment.
Latest papers
150 papers16 papers in the last four weeks, down 11% on the four weeks before. 0.1% of all new papers.