Quantitative Ultrasound

Quantitative ultrasound (QUS) aims to extract quantitative information about tissue properties from ultrasound images, going beyond traditional qualitative interpretation. Current research heavily utilizes deep learning, employing architectures like U-Nets, transformers, and Bayesian neural networks to improve parameter estimation (e.g., speed of sound, speckle statistics), segmentation accuracy, and uncertainty quantification. This work addresses challenges such as noise reduction, limited training data, and the need for robust and efficient algorithms for applications ranging from improved diagnostic accuracy to automated measurements in point-of-care settings. The ultimate goal is to enhance the diagnostic capabilities and efficiency of ultrasound imaging across various medical specialties.