SPS-BISP TC Webinar: How to Merge Deep Learning Models with Physics for Fast Multidimensional MRI

Date: 18 April 2023
Time: 3:00 PM CET (Local time)
Presenter(s): Dr. Matthews Jacob


Matthews JacobMatthews Jacob Mathews Jacob is a professor in the Department of Electrical and Computer Engineering and is heading the Computational Biomedical Imaging Group (CBIG) at the University of Iowa. His research interests include image reconstruction, image analysis, and quantification in the context of magnetic resonance imaging. He received his Ph.D from the Biomedical Imaging Group at the Swiss Federal Institute of Technology, and was Beckman postdoctoral fellow at the University of Illinois at Urbana Champaign.

Dr. Jacob is the recipient of the CAREER award from the National Science Foundation in 2009, the Research Scholar Award from American Cancer Society in 2011, and the Faculty Excellence Award for Research from University of Iowa in 2021. He is currently the associate editor of the IEEE Transactions on Medical Imaging and has served as the associate editor of IEEE Transactions on Computational Imaging from 2016-20. He was the senior author on two best paper awards (2015 & 2021) and one best machine learning paper award (2019) from IEEE ISBI. He was the general chair of IEEE International Symposium on Biomedical Imaging, 2020. He was elected as a Fellow of the IEEE (2022) for contributions to computational biomedical imaging.


Model-based deep learning (MoDL) methods that combine imaging physics with convolutional neural networks have shown great promise in improving image quality and computational efficiency compared to compressed sensing (CS) based MRI methods. However, they face challenges such as lower robustness to input perturbations, vulnerability to model and data mis-fits, and do not have energy-based formulations like CS methods. In addition, MoDL methods have a high memory demand and often require large fully-sampled datasets for training. As a result, it can be challenging to apply MoDL directly to free-running multidimensional MRI sequences that encode spatial & temporal dimensions and multiple contrast mechanisms. This is a significant limitation since free-running sequences are more time-efficient and enable different views and contrasts for tissue visualization. To address these challenges, we will present novel MoDL formulations that improve robustness, reduce vulnerability to model and data mis-fits, and reduce memory and data demand during training. These advancements will be demonstrated in challenging neuro, cardiac, and lung MRI applications involving free-running MRI sequences.