Short Courses (Group 1)

This introductory course is intended to give students and those newly entering the field of ultrasound imaging an overview of medical ultrasound. We provide a framework for introducing the concepts of how an ultrasound system images the body and how it can be applied to different applications. First, we present how tissues interact with sound waves through a series of simulators. Second, we follow signals through an imaging system as they turn from electrical pulses in a beamformer and transducer into acoustic waves and back into electrical signals and finally into an image. Third , we demonstrate different types of imaging modes and their role in the visualization of three-dimensional objects. Fourth, we explain how an ultrasound research system can be reconfigured to serve different functions and applications. Finally we provide a brief overview of the different branches and specialized applications of ultrasound. Throughout the course, we demonstrate the principles by using special purpose interactive computer simulators as well as live experiments and imaging with an ultrasound research system. 

Over the past years, deep learning has established itself as a powerful tool across a broad spectrum of domains. While deep neural networks initially found applications in the computer vision community, they have quickly spread over medical imaging applications, ranging from image analysis and interpretation to - more recently - image formation and reconstruction. Deep learning is currently also rapidly gaining attention in the ultrasound community. This course will first cover the basic principles of deep learning, ranging from understanding the relevance of sequential nonlinear transformations for representation learning to log-likelihood based optimization of neural network parameters. Optimization aspects such as the impact of local minima and saddle points in the solution space will also be touched upon. We will then discuss the design of effective neural network architectures, including dedicated solutions that e.g. leverage signal structure through unfolding methods. The second part will focus on the wealth of opportunities that deep learning brings for ultrasound imaging. Beyond image-level classification and segmentation, we will discuss neural networks for front-end receive processing, including beamforming, clutter suppression, and advanced applications such as super-resolution imaging.  

The advent of ultrafast ultrasonic scanners is paving today the way to tremendous applications in medical Ultrasound. This course will present the basic principles of Ultrafast Imaging (plane or diverging wave imaging, parallel receive beamforming, coherent plane wave compounding, ...) and their implications in terms of resolution, contrast and frame rates. It will also explain the analogy such concept with optical holography. For our purposes, theoretical aspects and experimental validations will be highlighted. The course will also emphasize technological issues and system architecture constraints. Far beyond breaking technological barriers, this concept of ultrafast imaging is currently changing the paradigm of ultrasound imaging. The course will illustrate how this concept leads to breakthrough innovations in the field by revisiting Bmode, Doppler, tissue strain and nonlinear imaging. Many examples (Shear Wave Imaging, Ultrafast Doppler, fUltrasound, Ultrafast Contrast Imaging, ultrafast Ultrasound Localization Microscopy,...) will illustrate the potential of this paradigm shift in ultrasound imaging.  

Super-resolution imaging has the capacity to distinguish and map structures that are smaller than the classical limit, typically a fraction of the wavelength. For ultrasound imaging, this means exploring features, such as blood vessels, in the micrometric range deep inside tissue. At the end of this course, students should be able to understand and reproduce super-resolution ultrasound imaging experiments, from data acquisition to image reconstruction, and apply such knowledge in their specific fields. We first explore the fundamental aspects of imaging resolution in ultrasound. Various approaches to bypass the diffraction-limit with microbubbles and other agents are presented. More particularly, we discuss ultrasound localization microscopy, which has recently improved the resolution for vascular imaging by more than 10-fold. We present its various steps, including separation, localization and tracking, and compare different approaches. Specific elements such as temporal resolution, motion correction or volumetric imaging are considered. We then detail the applications of super-resolution ultrasound for brain, tumor, kidney, liver lymph nodes and peripheral vessels imaging, along with future perspectives in the clinical and preclinical context. The last part of the course will include hands-on image processing of in-silico and in-vivo open data with provided ultrasound localization microscopy algorithms.