A Novel Benchmark Data Set for Adult Stem Cell Tracking in Time-Lapse Microscopy

Abstract

Event Abstract Back to Event A Novel Benchmark Data Set for Adult Stem Cell Tracking in Time-Lapse Microscopy Amir Madany-Mamlouk1*, Tim Becker2 and Daniel Rapoport3 1 University of Lubeck, Institute for Neuro and Bioinformatics, Germany 2 Graduate School at University of Lubeck, Germany 3 Fraunhofer EMB, Germany In modern stem cell therapies -- for cancer treatment as well as for cardiac operations -- typically both types of stem cells (autologous or allogenic) are immediately transplanted after isolation. Especially the potential of growing these cells in culture is neglected this way, even though clinical cell farming would significantly enhance the scope of stem cell treatment. There is just no profound diagnostic and documentation system for clinical cell farming.To gather such knowledge we established a robust and complete cell detection in time-lapse microscopy to have a highly effective tracking of several hundreds to thousands cells per image. Doing so, cell populations can be traced for days up to weeks. Unfortunately, the fully-automated cell tracking poses some challenging problems, especially as every new cell culture tends to vary strongly for several degrees of freedom, even within the same cell types.In this work, we will present the following three results of our ongoing project: 1. Cell Detection and Tracking on Phase-Contrast Microscopy Time-Lapse Data Essential for a robust and reliable cell tracking is a successful cell detection task, in which we find all cells in a given microscopy snapshot. We propose a preprocessing using a cascade of morphological filters, and a prototyping approach for unknown cell types using a supervised neural network. For certain cell types, the presented framework is capable of detecting up to 95% of all cells in real data scenarios and track most of them over all frames of the recording.(See the below picture, in which the red cell is of interest. Red dots mark the path this cell moved over time, yellow cells are sister cells and yellow dots their path over time) 2. Proposing a Benchmark Set for the Evaluation of Cell Tracking/Detection Methods. To evaluate our above approaches, we are currently building up a huge database of several thousand completely labeled cell culture images taken under different conditions and objectives. This data is collected in a semi-automatic fashion, with an initial fully-automated detection and an assisted manual correction and control step afterwards. The human operator has not to bother about all clear cases and can focus on the remaining 5% of cells, which is significant speaking of 500 instead of 10.000 cells to be hand-labeled. This benchmark database will be available to the public in the near future and might help to compare existing approaches in cell tracking.3. sensor fusion of phase contrast, light microscopy and immunofluorescence images.In a next step, we want to propose an approach to fuse all available information during cell tracking to gain insides of the cells behavior that is not visible observing a single measure alone. In detail, we will discuss the fusion of phase contrast and light microscopy images together with information about the optical flow in the image sequences and correlating this information to a standard measure, the immunofluorescence staining of the cells, for which cells typically have to be destroyed. INCF-09-106 Conference: Neuroinformatics 2009, Pilsen, Czechia, 6 Sep - 8 Sep, 2009. Presentation Type: Poster Presentation Topic: General neuroinformatics Citation: Madany-Mamlouk A, Becker T and Rapoport D (2019). A Novel Benchmark Data Set for Adult Stem Cell Tracking in Time-Lapse Microscopy. Front. Neuroinform. Conference Abstract: Neuroinformatics 2009. doi: 10.3389/conf.neuro.11.2009.08.081 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 22 May 2009; Published Online: 09 May 2019. * Correspondence: Amir Madany-Mamlouk, University of Lubeck, Institute for Neuro and Bioinformatics, Lubeck, Germany, madany@inb.uni-luebeck.de Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Amir Madany-Mamlouk Tim Becker Daniel Rapoport Google Amir Madany-Mamlouk Tim Becker Daniel Rapoport Google Scholar Amir Madany-Mamlouk Tim Becker Daniel Rapoport PubMed Amir Madany-Mamlouk Tim Becker Daniel Rapoport Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.