Machine learning in medical ultrasound to assist clinical diagnosis

摘要

I describe recent progress in machine learning in medical ultrasound imaging. At a high level, the two technologies appear to be well-matched as the former attempts to make sense of complex patterns, and the latter can generate images of complex acoustic patterns. Thus, the general challenge to the image analysis practitioner is to design machine learning architectures that model the patterns in real world ultrasound images to answer useful clinical questions.