27th ISMRM Annual Meeting • 11-16 May 2019 • Montréal, QC, Canada

Weekend Educational Session
Humans Learning to Do Machine Learning Right

Back to Program-at-a-Glance

Humans Learning to Do Machine Learning Right
Weekend Course

ORGANIZERS: Demian Wassermann, Matthias Guenther

Saturday, 11 May 2019
Room 710A  08:00 - 12:00 Moderators:  Demian Wassermann

Skill Level: Basic to Advanced

Session Number: WE-08

Introduction into the basics and experimental design of machine learning applications.

The class will present a mixed array of tutorials. There will be a focus classical and deep neural networks-based machine learning with applications to a diverse array of MRI applications. Moreover, bleeding-edge technologies and open questions in machine learning will be introduced.

Specific attention to minimum requirements in terms of experimental design and validation to ensure reproducibility of results obtained through machine learning techniques.

Target Audience
People using or planning to use machine learning for solving magnetic resonance in medicine problems.

Educational Objectives
As a result of attending this course, participants should be able to:
- Identify main current trends in machine learning;
- Assess machine learning literature with a more grounded knowledge; and
- Demonstrate basic understanding of experimental design and reporting of machine learning results.


  Best practices and pitfalls in applying Machine Learning to Magnetic Resonance Imaging
Bertrand Thirion
  Deep Neural Network Applications to Magnetic Resonance in Medicine
Tal Arbel
  Applications for Machine Learning in Medical Imaging
Polina Golland
  Experimental Design for Applications of Machine Learning in Magnetic Resonance in Medicine
Jean Baptiste Poline
  Break & Meet the Teachers
  State of the Art and Current Problems in Deep Learning
Daniel Rueckert
We will give an overview of the current state-of-the-art in deep learning for medical imaging applications such as reconstruction, segmentation and classification. In particular, we will illustrate deep learning approaches based on Convolutional Neural Networks (CNN). We will focus on deep learning models taht use encoder-decoder networks and show these can be used for tasks such as image reconstruction and image segmentation. We show some applications of CNNs in the context of image classification. Finally, we will discuss some open challenges for deep learning approaches such as explainability and verification of deep learning.

  Learning Image Reconstruction: AUTOMAP
Bo Zhu
In this educational talk we describe AUTOMAP, a generalized image reconstruction method utilizing machine learning end-to-end from raw k-space to the final image, enabling reconstruction of arbitrary spatial encoding schemes and encoding spaces (not limited to Fourier) and also featuring noise-robustness which is produced during the training process.

  Machine Learning Applications to Diffusion MRI Microstructure
Marco Palombo
Diffusion MRI (dMRI) signal is sensitive to the tissue architecture at the microscopic scale. Modern machine learning and deep learning techniques can be used to learn the mapping between acquired dMRI signal and specific features of the tissue microstructure. However, experimental design and validation of training sets are essential for reliable supervised and semi-supervised learning and reproducibility and uncertainty of prediction are still open questions. This lecture provides the key concepts behind machine learning applications to dMRI signal analysis for tissue microstructure quantification and show the audience various techniques which have been recently used.  

  Lunch & Meet the Teachers
Back to Program-at-a-Glance   |   Back to Top
The International Society for Magnetic Resonance in Medicine is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.