Conference Sessions
Data Handling and Analysis (DHA)

DHA.1: Machine Learning for Analysis and Interpretation of Microscopy Imaging Data

Machine learning methods such as deep learning have revolutionized applications in image analysis and understanding during the last decade. The developments have also impacted microscopy image analysis and processing where the most successful methods in object segmentation and detection, image reconstruction, classification and denoising are nowadays taking advantage of machine  learning techniques. With the aid of open datasets, machine learning solutions continues to develop towards easy-to-use platforms that include pre-trained models for various applications whereas research challenges focuses on unsupervised and weakly supervised learning.

This session presents the latest advancements in methods and software utilizing machine learning applied broadly in different microscopy modalities e.g. light, electron and X-ray microscopy and digital pathology. We also invite studies describing open datasets to be presented in this session. In addition, the session hosts discussion on the challenges as well as future perspectives that the machine learning currently holds.

Related Conference Theme/s: Artificial Intelligence in Big Data Analysis and Computational Microscopy

Session Chairs and Invited Speakers
  • Session Chairs

      Lassi Paavolainen (Ph.D.) is an Academy of Finland postdoctoral researcher at the Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki. With a background in computer science, Dr. Paavolainen has worked in collaboration with biologists over a decade to solve microscopy image analysis questions. During his Ph.D. studies, his work included image analysis of fluorescence and electron microscopy imaging data and development of widely used open source BioImageXD software for analysis and visualization of multi-dimensional microscopy image data. Since 2015 Dr. Paavolainen has worked as a postdoctoral researcher at FIMM in Horvath and Kallioniemi research groups. His research focuses on analysis of large-scale high-content imaging data in collaborative precision medicine research projects studying patient-derived cancer cells. During these projects he contributed to the development of open source Advanced Cell Classifier software for phenotypic classification of cell. Dr. Paavolainen was one of the founders and the first head of FIMM High-Content Imaging and Analysis unit. Currently he is a principal investigator of Academy of Finland funded project Deep learning for phenotypic profiling of cancer cells. He is a representative of Finland in Network of European Bioimage Analysts (NEUBIAS) COST project and an active participant in NEUBIAS WG5 activities. In addition, he is the vice president of CytoData Society that he co-founded, and the information officer in the board of Nordic Microscopy Society. His current research interests are in solving various image analysis questions using deep learning especially with limited amount of annotated data.


      Riku Turkki, PhD, is a post doctoral researcher at the Science for Life Laboratory (SciLifeLab) and Karolinska Institutet in a research group lead by Prof. Olli Kallioniemi. He holds a researcher position at the Helsinki University Hospital (HUS) and a visiting scientist position at the Institute for Molecular Medicine Finland (FIMM). Dr. Turkki received his MSc in Electrical Engineering at the University of Oulu and PhD in Medicine at the University of Helsinki. Dr. Turkki's main interest is to apply and develop machine learning methods to better understand medical imaging data, strongly focusing on applications in precision medicine with translational aspect. During his thesis, Dr. Turkki studied utilization of computer vision methods for digital pathology in cancer research. In his post doctoral research, Dr. Turkki is concentrating on bringing the latest machine learning, and especially deep learning methods, in the analysis of digital pathology as well as high content imaging. He is also working with unsupervised learning and weakly supervised learning method development.

  • Invited Speakers

      Talk Title: TBC

      Anna Kreshuk joined EMBL in July 2018 as a Group Leader in the Cell Biology and Biophysics Unit. Her research focuses on machine learning-based methods for the analysis of biological images. Right now, she is especially interested in large-scale image and volume segmentation. Besides, Anna is leading the development of the ilastik software, aiming to make such methods available to life scientists without computational expertise. Previously, she was a PostDoc at the Heidelberg Collaboratory for Image Processing and a visiting scientist at the HHMI Janelia Research Campus. Anna holds a PhD in Computer Science from the University of Heidelberg and a Diploma in Mathematics from Lomonosov Moscow State University. In between the two degrees, she worked at CERN in Geneva as a scientific programmer for the ROOT framework. 


        Talk Title: Improving image analysis and machine learning using a novel image representation

      Ivo Sbalzarini is the Chair of Scientific Computing for Systems Biology on the faculty of computer science of TU Dresden and a director of the Center for Systems Biology Dresden. He also is a Senior Research Group Leader with the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden, Germany. He completed his doctorate in computer science in 2006 at ETH Zurich, Switzerland (Chorafas Award, Weizmann Institute of Science), where he formed a close collaboration between biology and computer science. After a research stay at the California Institute of Technology (Caltech, Pasadena, USA), he was named Assistant Professor for Computational Science in the Department of Computer Science of ETH Zurich. In 2012, Ivo and his group moved to Dresden, where he became one of the founding members of the Center for Systems Biology. He also serves as a research avenue leader of the Federal Cluster of Excellence “Physics of Life", Dean of the International Max Planck Research School in Cell, Developmental, and Systems Biology, and Vice-Dean of the Faculty of Computer Science. Ivo’s research focuses on developing theories, algorithms, and computational methods that help advance the life sciences. This includes computer simulation methods for biological systems and processes, bioimage analysis and bioimage computing, virtual and augmented reality technology for microscopy, machine learning approaches for biology, and applications in the systems biology of embryo development. 

    • PETER HORVATH (Institute of Biochemistry Hungarian Academy of Sciences, HU)

      Talk Title: Life beyond the pixels: machine learning and image analysis methods for light microscopy

      Peter Horvath (1980) is currently a director and group leader in the Biological Research Center in Szeged and holds a Finnish Distinguished Professor (FiDiPro) Fellow position in the Institute for Molecular Medicine Finland (FIMM), Helsinki. He graduated as a software engineer and mathematician, and received his Ph.D. from INRIA and University of Nice, Sophia Antipois, France in satellite image analysis. Between 2007 and 2013 he was a senior scientist at the ETH Zurich, in the Light Microscopy Centre. Peter Horvath is interested in solving computational cell biology problems related to light microscopy and is involved in three main research fields; 2/3D biological image segmentation and tracking; development of microscopic image correction techniques; machine learning methods applied in high-throughput microscopy. He is the co-founder of the European Cell-based Assays Interest Group and the councillor of the Society of Biomolecular Imaging and Informatics.

    • Geert Litjens (Radbound University Medical Center, NL)

      Talk Title: Deep learning in digital pathology and microscopy for research and clinical practice

      Geert Litjens is an assistant professor of Computational Pathology at the Radboud University Medical Center. His research is at the intersection of machine learning, medical imaging, and oncology. He co-chairs the Computational Pathology Group, which develops automated machine learning systems for cancer detection, biomarker discovery and quantification, and improved prognostication. He is also the developer of the ASAP software package for analyzing and visualizing whole-slide images and (co-)organizer of several high-profile challenges in medical imaging such as PROMISE12 and CAMELYON. He (co-)authored over 60 publications in medical, imaging, and machine learning conferences and journals.

DHA.2: Advances in 3-Dimensional Image Reconstruction

The session will cover the quantitative analysis of X-ray and electron based 3D microscopy data. Electron and X-ray 3D microscopy is widely used in material science and is an emerging field with great potential in bioimaging. With the possibility of imaging 3D structures at high spatial or temporal resolution both ex and in vivo allows structural quantification that has many applications in bioimaging. In situ 3D imaging within materials science and 3D imaging of the interface between biology and materials science often requires imaging at different length scales. The full potential of electron and X-ray 3D microscopy can only be utilized having automated analysis methods and subsequent modeling tools. This session will focus on recent developments within X-ray 3D microscopy for both laboratory and large-scale synchrotron facilities as well as applications of 3D imaging techniques by electron sources. Both destructive (slice and view) and non-destructive 3D imaging applications will be discussed.

Both destructive (slice and view) and non-destructive 3D imaging applications will be discussed.

Related Conference Theme/s: Artificial Intelligence in Big Data Analysis and Computational Microscopy

Session Chairs and Invited Speakers
  • Session Chairs

      The research of Sara Bals focuses on electron tomography for reconstructing 3D morphologies of nanostructures, even with atomic resolution. She is currently a Full Professor at the Department of Physics, University of Antwerp and she received her Ph.D. degree (2003) from the same University. She did post-doctoral work at the National Centre for Electron Microscopy at the Lawrence Berkeley National Laboratory in California (USA). Sara Bals is an expert in the field of electron microscopy and electron tomography applied to functional nanomaterials. She is the author of more than 300 ISI contributions, including Nature or Science type contributions. Her work has been cited more than 8000 times and she has an h-index of 48. Over the last 15 years, she gave more than 40 invited presentations at international conferences and workshops. She is the co-organizer of the bi-annual EMAT Workshop on Electron Microscopy, was the co-organizer of an EMRS Fall Meeting symposium and has chaired several sessions at international microscopy conferences. In 2013, she received an ERC Starting Grant concerning 3D characterization of nanostructures by electron tomography (Colouratom). In 2015, she reached the finals of the New Scientist Talent of the Year Election. In 2016, she became “Laureate of the Academy for Natural Sciences” awarded by the Royal Flemish Academy of Science. She is currently a Francqui Research Professor and her ERC Consolidator Grant (Realnano) started in May 2019.




    • Sandra Van Aert (University of Antwerp, BE)

      Sandra Van Aert received her Ph.D. at the Delft University of Technology (The Netherlands) in 2003. Thereafter, she joined the Electron Microscopy for Materials Research (EMAT) group of the University of Antwerp (Belgium) where she is currently a Full Professor. Her research focuses on new developments in the field of model-based electron microscopy aiming at precise measurements of structure parameters. Model-based microscopy allows one to measure 2D atomic column positions with subpicometer precision, to measure compositional changes at interfaces, to count atoms in an atomic column with single atom sensitivity, to unscramble mixtures of elements, and to reconstruct 3D structures with atomic resolution. She received the 2011 European Microscopy Society Outstanding Paper Award and the 2017 Ernst Ruska Prize for achievements on ‘New techniques for optimum quantitative analysis of electron microscopy data’. In 2018, she obtained an ERC Consolidator Grant entitled ‘Picometer metrology for light-element nanostructures: making every electron count’.

  • Invited Speakers

      Talk Title: Towards real-time, adaptive microscopy

      Joost Batenburg published more than 80 journal articles and more than 60 conference papers in the field of tomographic image processing and reconstruction. From 2013 till 2017 he chaired the EU COST Action EXTREMA on advanced X-ray tomography. He pioneered the field of discrete tomography, developing the first large-scale reconstruction methods. His current research focuses on creating a real-time tomography pipeline, funded by an NWO Vici grant. He is responsible for the FleX-Ray lab, where a custom-designed CT system is linked to advanced data processing and reconstruction algorithms.


      Talk Title: Automated electron Diffraction Tomography – advantages of STEM acquisition for crystal structure analysis of nanocrystals

      After studying chemistry at the universities of Kaiserslautern and Mainz, Institute of Physical Chemistry she started her PhD in 1994 in the Insitute of Inorganic chemistry. This provided the possibility to work in a variety of fields such as photon correlation spectroscopy, X-ray scattering on single crystals and powder, crystal growth, computer simulation, modeling, electron microscopy and electron crystallography. In 1997 she started her habilitation with the aim to develop a reliable method for structure analysis using electron diffraction. Since 2001 she is responsible for the "Centre for high resolution electron microscopy Mainz" (EMZ-M) at the University of Mainz and since 2012 she is additionally appointed as Professor for Electron Crystallography in Darmstadt. In 2007 the method for automated diffraction tomography (ADT), developed in her group, proved suitable to solve "ab initio" crystal structures from nano particles of a wide range of materials and developed until today into a widely recognized method on the brink of routine use. Throughout the years she was active for different associations like ECA(SIG4), IUCr(CEC) and is now chair of the National Committee of the DGK. Additionally, she was teaching in international courses and organized Schools on Electron crystallography like in 2011 in Erice and 2014 in Darmstadt.


DHA.3: Machine Assisted Acquisition and Analysis of Microscopy Data

Cutting-edge microscopy does not only require state-of-the-art instruments and detectors but also innovative approaches and programs for machine-assisted collection and analysis of data. Recent advances in instrumentation and computing capabilities enable the application of machine learning for the processing of microscopy datasets during and after acquisition. Furthermore, the development of sustainable, open-source and user-friendly software is of paramount importance to make these algorithms and workflows available widely in the scientific community and to promote reproducible research. This symposium will feature recent progress in data acquisition scheme, processing workflow, algorithm and software in microscopy with a focus on, but not limited to, open-source software and machine learning.

Related Conference Theme/s: Artificial Intelligence in Big Data Analysis and Computational Microscopy

Session Chairs and Invited Speakers
  • Session Chairs

      Daniel Baum is head of the working group »Image Analysis in Biology and Materials Science« at the Zuse Institute Berlin (ZIB). He studied computer science at the Humboldt-Universität zu Berlin and the University of Edinburgh. In 2007, he received his Dr. rer. nat. degree from the Freie Universität Berlin, specializing in comparative molecular surface analysis. His current research interests are in the fields of visualization and data analysis of molecular structures as well as all kinds of image data of specimens from materials science and biology. Methodologically, his interests include image segmentation, feature extraction, the geometric reconstruction of features, and shape analysis. His focus is on interdisciplinary work with the aim of bringing practical, automated solutions to other research communities that enable researchers in these fields to make their work more efficient and to investigate statistically relevant amounts of data.


        After a first degree in Material Science, Lewys received his PhD from the Department of Materials at the University of Oxford in 2013. This focussed on two themes; scanning stability in the aberration-corrected scanning transmission electron microscope (AC-STEM) and applications of focal series of annular dark-field data. After 4 years as a post-doc in the Nellist group at David Cockayne Centre for Electron Microscopy, in 2017 Lewys moved to Trinity College Dublin to found the Ultramicroscopy Group as the new Ussher Assistant Professor in Ultramicroscopy. In 2019, Lewys was awarded a Royal Society & Science Foundation Ireland (SFI) Universty Research Fellowship to expand the hardware instrumentation development activities of the group. Lewys has authored around 100 articles and proceedings with more than 1000 combined citations, and has launched two commercial software plug-ins for Digital Micrograph in collaboration with HREM Research. Lewys is a co-director of the SFI-EPSRC Centre for Doctoral Training in Advanced Characterisation, an Associate Editor of the journal Advanced Structural and Chemical Imaging, and has been a Fellow of the Royal Microscopical Society since 2015.

  • Invited Speakers

      Talk Title: Light-sheets, Neural Networks and Zebrafish - Science at the Interface

      Royer first studied engineering in his native France and then obtained a master's degree in Artificial Intelligence, specializing in Cognitive Robotics, followed by a Ph.D. in Bioinformatics from the Dresden University of Technology in Germany. He then joined Gene Myers’ lab, first at HHMI's Janelia Farms and then at the Max Planck Institute of Molecular Cell Biology and Genetics, where he developed novel technology at the intersection of computer science and microscopy, including the first adaptive multi-view light-sheet microscope, which he developed in collaboration with Philipp Keller. As a group leader at CZ Biohub, Royer and his team are building ‘discovery machines’ that not only acquire image data, but also perform online processing, instant 3D visualization, adaptive imaging, and automated photo-manipulation. These integrated instruments bring together state of the art optics, robotics, machine learning, and image analysis with the aim of advancing beyond the automation of repetitive tasks and into the realm of actual automated scientific reasoning.


      Talk Title: Analytical electron tomography: tools for achieving fast and reliable reconstructions

      Zineb Saghi graduated from INSA Lyon where she obtained a master’s degree in Electrical Engineering and a master of research in Medical Imaging. She received her Ph.D. degree (2009) from the University of Sheffield, under the supervision of Gunter Moebus. Her work aimed at developing electron tomography for the 3D analysis of nanomaterials, and exploring its prospects at atomic resolution using aberration-corrected microscopes. She then joined Paul Midgley’s group at the University of Cambridge, where she worked on compressed sensing and its application to electron tomography in biology and materials science.  She also explored subsampling and inpainting approaches to speed up the acquisition process. Since 2015, she has a position of research scientist at CEA-Leti in Grenoble, France, working on advanced characterization of semiconductor devices using EELS and EDS tomography. Her research interests focus on the application of high speed/low-dose acquisition schemes, machine learning and compressed sensing techniques to analytical electron tomography.


Share this