From shallow to deep learning for inverse imaging problems: some recent approaches

Add to your list(s) Download to your calendar using vCal

• Prof. Carola-Bibiane Schönlieb, Cambridge University
• Wednesday 25 September 2019, 14:00-15:00
• Room 611, EEE.

If you have a question about this talk, please contact Vanessa Rodriguez-Gonzalez.

Abstract: In this talk we discuss the idea of data-driven regularisers for inverse imaging problems. We are in particular interested in the combination of model-based and purely data-driven inversion approaches. In this context we will make a journey from “shallow” learning for computing optimal parameters for variational regularisation models by bilevel optimization to the investigation of different approaches that use deep neural networks for solving inverse imaging problems. The talk is furnished with application of these ideas to medical imaging, in particular computed tomography.

Bio: Carola-Bibiane Schönlieb is an Austrian mathematician who works in image processing and partial differential equations. She is a Fellow of Jesus College, Cambridge and a Professor in Applied and Computational Analysis in the Department of Applied Mathematics and Theoretical Physics at the University of Cambridge. She is the author of the book Partial Differential Equation Methods for Image Inpainting (Cambridge University Press, 2015), on methods for using the solutions to partial differential equations to fill in gaps in digital images.

Schönlieb earned a master’s degree in mathematics from the University of Salzburg in 2004. She completed her Ph.D. at Cambridge in 2009. Her dissertation, Modern PDE Techniques for Image Inpainting, was supervised by Peter Markowich. After postdoctoral study at the University of Göttingen she returned to Cambridge as a lecturer in 2010.

In 2016 Schönlieb won the Whitehead Prize of the London Mathematical Society “for her spectacular contributions to the mathematics of image analysis”. She won a Philip Leverhulme Prize in 2017, and is the 2018 Mary Cartwright Lecturer of the London Mathematical Society. Since 2016 she has also been a Fellow of the Alan Turing Institute. Carola is Director of the Cantab Capital Institute for the Mathematics of Information and she is Co-Director of the Centre for Mathematical Imaging in Healthcare—both are based at the University of Cambridge.

This talk is part of the Vanessa Rodriguez-Gonzalez's list series.

This talk is included in these lists:

• Room 611, EEE

Note that ex-directory lists are not shown.