Carola-Bibiane Schönlieb : Inverse Problems in Imaging: From Differential Equations to Deep Learning скачать в хорошем качестве

Carola-Bibiane Schönlieb : Inverse Problems in Imaging: From Differential Equations to Deep Learning

Recording of Carola-Bibiane Schönlieb’s (University of Cambridge) talk on May 12, 2022, at the EPFL Seminar Series in Imaging. Abstract. In the last couple of years the processing and analysis of imaging data has undergone a significant paradigm shift, from knowledge driven approaches that derive imaging models from first principles to purely data driven approaches that derive models from data. In this talk I will discuss image processing methods that operate at the interface of these paradigms and feature both a knowledge driven (mathematical modelling) and a data driven (machine learning) component. Mathematical modelling is useful in the presence of prior information about the imaging data and relevant features of interest, for narrowing down the search space, for highly generalizable methods with solutions that come with theoretical solution guarantees. Machine learning on the other hand is a powerful tool for customising image processing methods to individual data sets. Their combination is the topic of this talk, furnished with examples for image classification under minimal supervision with an application to chest x-rays, tomographic image reconstruction with learned priors and fast spatio-temporal MRI. See upcoming talks on: imagingseminars.org The EPFL Seminar Series in Imaging is run by EPFL Center for Imaging: imaging.epfl.ch