skip to content

Researcher: Yury Korolev, and Carola-Bibiane Schönlieb

The goal of this project is to develop a data driven regularisation theory for inverse problems, extending classical, model based results to the model-free setting and developing novel data driven inversion algorithms with regularisation guarantees. We want to understand to what extent classical model based regularisation theory can be carried over to the purely data driven setting. A typical question that we ask ourselves is whether regularisation can be achieved by the training data alone, e.g., by properly choosing the training set or aligning the level of noise in the training data with that in the measurement. Using this understanding, we aim to develop novel data driven regularisation methods and apply them in inverse imaging problems.


Related Publications 

Data driven regularization by projection
A Aspri, Y Korolev, O Scherzer – Inverse Problems (2020) 36, 125009
Data Driven Reconstruction Using Frames and Riesz Bases
A Aspri, L Frischauf, Y Korolev, O Scherzer (2021)303