CellPhaser – An intuitive machine learning solution for the analysis of holographic cell images

Name of the participant: Stefan Röhrl

Description of the IT research project: The analysis of cellular structures gives us deep insights into the health status of patients, which is why they make up a large part of medical laboratory tests. Digital holographic microscopy (DHM) allows label-free analysis of cells near in vivo conditions without the time-consuming sample preparation of conventional methods. In contrast to light microscopy, DHM is used to quantitatively measure the phase shift caused by cells, which correlates directly with the internal condition of the cells. This allows differentiation and morphological analysis of cells without the need for staining or labelling. To enable clinically relevant diagnostics with DHM, a Machine Learning Toolbox for the analysis of different clinical questions will be developed in this project. For an application in the clinical environment the found algorithms will be integrated into an ergonomic user interface called “CellPhaser”. The CellPhaser functions as the software core of the “CellFace” project at the Central Institute for Translational Cancer Research (TranslaTUM), with the goal of establishing DHM-based, label-free flow cytometry as a platform technology for cellular diagnostics.

Software Campus partner: TU Munich, Merck

Implementation period: 1.1.2020 – 30.09.2021