Spotlight on research Irene Yu-Hua Gu

In this talk, we describe several deep learning methods we have been studying, based on MR images, that show good potentials for computer-assisted diagnosis of neurological diseases (e.g., detecting Alzheimer’s disease, predicting molecular subtypes of brain tumors without biopsy)


4th of Jun 2021, 13.00 (Swedish time)
Online, Zoom

 Register here

Title: Deep Learning Methods for Assisting Medical Diagnosis of Neurological Diseases: methods, possibilities and challenges


Recent developments in imaging technology have revolutionized healthcare. Medical images are widely used for diagnosis and therapy monitoring with significant impact on patient treatment outcome. Despite these advances, routine clinical MRI data interpretation is still performed mostly by human experts. 

Despite the impressive progress of AI technologies especially in computer vision, AI technologies/deep learning-assisted medical diagnosis for neurological diseases remains an open and challenged research area.
In this talk, we describe several deep learning methods we have been studying, based on MR images, that show good potentials for computer-assisted diagnosis of neurological diseases (e.g., detecting Alzheimer’s disease, predicting molecular subtypes of brain tumors without biopsy). We also discuss some common challenges when dealing with medical datasets, e.g., the existing of several small datasets each consisting of a small number of patients, the collection of patient data that is incomplete, or only part of training data is annotated, these issues may impact the training process in deep learning, leading to low generalization performance on the test data. 

We then describe several possible solutions to these challenging issues, including domain mapping for combining several small training datasets, MRI data augmentation for missing data and for synthetic patients’ data. Finally, we demonstrate that these methods can be rather effective and hence offer great potentials for further research development in this area.


MicrosoftTeams-image (3).png


About the speaker: Irene Yu-Hua Gu received Ph.D. degree in electrical engineering from Eindhoven University of Technology (The Netherlands), in 1992. From 1992 to 1996, she was Research Fellow at Philips Research Institute IPO, (The Netherlands), post dr. at Staffordshire University (U.K), and Lecturer at the University of Birmingham (U.K). Since 1996, she has been with the Department of Electrical Engineering (former name: Department of Signals and Systems), Chalmers University of Technology, Sweden, where she became a professor (bitr. professor) in 2004, and a full professor since 2008. Her research interests include statistical image and video processing, video object tracking and recognition, machine learning and deep learning, and signal processing with applications. During the last several years her main research has been focused on biomedical image analysis and deep learning. Dr. Gu served as Associate Editor for IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans, and Part B: Cybernetics (2000-2005), Associate Editor for EURASIP Journal on Advances in Signal Processing (2005 – 2016), Editorial Board of the Journal of Ambient Intelligence and Smart Environments (2011-2019), Chair of the IEEE Swedish Signal Processing Chapter (2001-2004). She is a senior member of IEEE and currently serves as a Senior Area Editor (SAE) for the IEEE Signal Processing Letters. She has coauthored 200+ papers, and has been in the Guide2Research ranking list on “Top Computer Science and Electronics in Sweden” 2020 and 2021.

Category Seminar; Event
Location: Online
Starts: 04 June, 2021, 13:00
Ends: 04 June, 2021, 13:45

Page manager Published: Mon 24 May 2021.