Name: Tamas Ungi
Affiliation: School of Computing, Queen’s University, Kingston, ON, Canada
Primary research interest: Ultrasound image processing and guided medical interventions
Title of the lecture: Ultrasound image processing with artificial neural networks
Keywords: sonography, deep learning, image-guided therapy
Summary: Ultrasound scanners are becoming smaller and more affordable. Some predict that ultrasound will become the stethoscope of the future, as healthcare workers will be able to carry them in their pockets and use them routinely in patient assessments. But ultrasound is the most challenging medical imaging modality to interpret. The difficulty in steering ultrasound beams and the presence of acoustic artifacts result in images with highly variable quality and patterns. Recent advances in artificial intelligence may help by automatically providing annotations for ultrasound images. In this lecture, we will review the specifics of ultrasound imaging and common image processing methods that simplify interpretation. We will explore real-time ultrasound segmentation methods and their practical applications, particularly in image-guided surgery.