UCL Centre for Medical Image Computing


Pascal Fernsel - CMIC seminar series

13 March 2019, 4:00 pm–5:00 pm

Pascal Fernsel - CMIC seminar series

Event Information

Open to





Roberts 421
Roberts Building
Malet Place

Pascal Fernsel, University of Bremen


Title: Nonnegative Matrix and Tensor Factorizations – Theory and Selected Applications



Matrix and tensor factorization methods play an essential role in gaining lower dimensional approximate representations of datasets and have been widely used in machine learning and big data applications as a tool for model reduction, feature extraction, blind source separation and noise removal. In this talk, we focus on the particular case of nonnegative factorization approaches, which are favourable for a range of applications where the data under investigation naturally satisfies a nonnegativity constraint.

On the theoretical side, we will briefly discuss the basics of nonnegative matrix and tensor factorization (NMF, NTF) approaches, describe possible solution techniques while focussing on majorize-minimization algorithms for NMF and talk about connections between orthogonal NMF and K-Means methods.

Furthermore, we will look into selected applications in the field of medical imaging such as tumor typing and classification tasks in pathology for MALDI imaging datasets and possible applications of nonnegative tensor factorization methods to MRI.



Pascal Fernsel is a Ph.D. student and member of the Research Training Group Pi³ at the Center for Industrial Mathematics (ZeTeM) at the University of Bremen, Germany. His current research interests include machine learning together with matrix and tensor factorization methods, cluster analysis and signal processing.