Dr. Leonid Kostrykin is a postdoctral researcher at Heidelberg University. His main interests are globally optimal methods for computer vision (also image processing and computer graphics). As the ELIXIR Germany Officer of the Heidelberg Center for Human Bioinformatics (HD-HuB), he drives new developments of image analysis tools for Galaxy Europe and Bioconda.

Projects

Microscopy Image Analysis Course 2023, Organization and Teaching.
We gave an introduction into the field of microscopy image analysis using automated software tools (ImageJ and Galaxy). The course was attended by 17 participants, including scientists from Germany and Italy.
- uni-heidelberg.de
- denbi.de
Contributing Tools to Galaxy Europe and Bioconda (2022–present).
We aim to constantly contribute state-of-the-art image analysis tools to the Galaxy web platform and Conda ecosystem to facilitate further research and make the methods accessible to end-users.
- Bioconda Recipes
- Galaxy Image Analysis
Image Analysis for Visual Inspection of Electric Generators (2019–present).
We are concerned with the development of methods for large-scale image stitching of video data, acquired by small robots under low-light conditions to facilitate the inspection process of electric generators.
- Kostrykin et al. (ICCAS 2021)
- Patent WO 2023/025570 A1
Globally Optimal Segmentation using Shape and Intensity Information (2016–present).
We develop new globally optimal model-based methods for segmentation of cell nuclei. Our latest method is SuperDSM, which leverages superadditivity and deformable shape models.

Affiliations

Biomedical Computer Vision Group (BMCV), Heidelberg University,
- Research Fellow, PhD Student (2016–2022),
- Postdoctoral Researcher (2022–present).
- PhD Thesis: Globally Optimal Segmentation using Shape and Intensity Information
- BMCV
- Dissertation
Forschungszentrum Jülich GmbH,
- ELIXIR Germany (2023).
Heinrich-Heine-Universität Düsseldorf,
- Student of Computer Science (2013–2016).
- M.Sc. Project: Blind Deconvolution of Noisy Image Sequences
- Kostrykin and Harmeling (2022)
MediTEC, RWTH Aachen University,
- Student Assistant (2013–2015).
- RWTHmediTEC/Carna
University of Applied Sciences Aachen,
- Student of Scientific Programming (2010–2013).
- B.Sc. Thesis: Virtual Coronary Angiography based on CT Data
- RWTHmediTEC/DICOMViewer

Talks and Posters

Talk. Galaxy Community Conference: Image Analysis in Galaxy. Jun 2024.
Talk. IEEE International Symposium on Biomedical Imaging: Robust Graph Pruning for Efficient Segmentation and Cluster Splitting of Cell Nuclei using Deformable Shape Models. May 2024.
Invited talk. Next Generation Bioimage Analysis Workflows Hackathon: Workflows for Image Analysis using Galaxy. Irchel Campus, University of Zürich, Nov 2023.
Poster. DAGM German Conference on Pattern Recognition: Superadditivity and Convex Optimization for Globally Optimal Cell Segmentation using Deformable Shape Models. Sep 2023.
Talk. International Conference on Control, Automation and Systems: Globally Optimal and Scalable Video Image Stitching for Robotic Inspection of Electric Generators. Oct 2021.
Talk. IEEE International Symposium on Biomedical Imaging: Segmentation of Cell Nuclei using Intensity-based Model Fitting and Sequential Convex Programming. Apr 2018.

Mentorship and Teaching

Organization of (2023) and Teaching within (2019, 2023) the de.NBI training course “Microscopy Image Analysis 2023”, de.NBI Project, HD-HuB.
- uni-heidelberg.de (2019)
- uni-heidelberg.de (2023)
Annual supervision of seminar projects (2022) on large-scale data analysis using Python.
- hdsu.org
Supervision of student projects:
- J. Jegelka, “Interpretation of Machine Learning Models for Classification of Cell Nuclei and Imaging Artifacts” (2020–2021).
- S. Kopetschke, “Comparison of Different Image Stitching Software and Methods for Microscopy Images” (2019).
- R. Rappold, “Fluorescence Microscopy Image Data Search, Hyperparameter Optimization, and Comparison of Standard Methods for Cell Nuclei Segmentation” (2018).
Annual organization (2020–present) and supervision (2017–present) of programming classes using Java (2017–2019) and Python (2020–present) for undergraduate students.