On tomographic imaging with limited data

Title

On tomographic imaging with limited data

Speaker

Tatiana Bubba - Università degli Studi di Ferrara

Abstract

Tomographic imaging is an essential non-destructive technique in medicine and industry for visualizing internal structures of objects. A key challenge lies in the inherent noise and limitations (scarcity or undersampling) of measurements, requiring accurate modeling and the integration of prior information for reliable reconstructions. Consequently, the challenge of limited data tomography has spurred considerable theoretical and numerical investigation in recent years, with significant attention towards, inter alia, variational regularization and data-driven approaches. This talk will highlight applications of limited data tomography involving both static and dynamic targets. In these scenarios, the inherent ill-posedness is tackled by leveraging the synergistic combination of sparse regularization theory, applied harmonic analysis, microlocal analysis tools, and (self-)supervised learning.

This talk is based on a series of joint works with T. Heikkilä, D. Labate, M. Lassas, S. Mukherjee, L. Ratti, M. Santacesaria, A. Sebastiani and S. Siltanen.

Bio

Tatiana Bubba is an Assistant Professor in Numerical Analysis at the University of Ferrara, Italy. After obtaining her PhD in 2016 from the same university, she became postdoctoral researcher at the University of Helsinki, Finland, whereshe was Academy Postdoc from 2020. In 2021 she relocated to the UK, with a Royal Society Newton International Fellowship at the University of Cambridge, continuing as a Lecturer at the University of Bath until 2024. Her interest lies in computational inverse problems, especially in tomographic applications. Currently, her research revolves around combining sparse regularisation techniques based on multiresolution system like shearlets, with deep learning approaches.

When

Thursday, July 10th, 12:00

Where

Room 322, UniGe DIBRIS/DIMA, Via Dodecaneso 35