Machine Learning Applications in Medicine - Is There Still Room for Signal Processing Innovations?
Mert Sabuncu
December 14, 2022, Wednesday, 3:00 PM - 4:00 PM EST
Over the last decade, deep learning has transformed biomedical imaging, including brain imaging, from enhancing acquisition to maximizing downstream utility of scans. My research group has been at the forefront of this revolution, developing novel methods that have laid the foundation for next-generation tools. As I will describe in my talk, much of this progress relies on predictive models and thus can be viewed as "curve-fitting" with general-purpose models. I will then show a couple of examples form our recent work, where we move beyond the curve-fitting paradigm and custom build models in ways that allow us to gain novel insights and understand how the output was computed. In particular, I will focus on image registration and neural decoding ("mind reading" from fMRI data) as two application domains.
Mert Sabuncu received a PhD degree in Electrical Engineering from Princeton University, where his dissertation work focused on the image processing problem of establishing spatial correspondence across multiple clinical scans. Mert then moved to Massachusetts Institute of Technology (MIT) to do a post-doc at the Computer Science and Artificial Intelligence Lab (CSAIL), where he worked with Polina Golland on biomedical image analysis.

The Sabuncu Lab is a research group spanning Cornell Tech and Weill Cornell Medicine Radiology. We conduct research in the field of biomedical data analysis, in particular imaging data, and with an application emphasis on neuroscience and neurology. We use tools from signal/image processing, probabilistic modeling, statistical inference, computer vision, computational geometry, graph theory, and machine learning to develop algorithms that allow us to learn from and exploit large-scale biomedical data.