Image Processing & Analysis
During my scientific carreer and my position as Leader of Image and Data Analysis Facility (9 years employment) at the German Center of Neurodegenerative diseases I have carried out, accompained or managed many projects in the field of quantitative image analysis. This includes
- image analysis training for scientists
- image restoration (dealing with measurement artifacts)
- automated detection
- classification of objects
- classification of images
- handling processing of large amounts of image data
I have strong experience in the analysis of biomedical image data, such as:
- fluorescence microscopy
- time series data
- histology, quantitative histology
- MRI and CT data
- electron microscopy
Tools for image processing:
- Python
- ImageJ
- CellProfiler
- Napari
- Imaris
Related Projects
-
Python Developer / Project Lead
2016
Development of YAPiC, an open source software for analyzing biomedical image data using deep learning.
Tasks:
- Project management
- Conception and algorithm development
- Presentation of the software at international conferences
- Deployment, development of CI/CD pipelines
- Management of the further development of the tool by the open source community
Tools:
PythonTensorFlowTravis-CIGitHub YAPiC Website -
Data Scientist / Image Processing Specialist
2017
Automated characterization of tissue samples using deep learning
As part of a medical research project, tissue samples were photographed with an automated microscope. Based on Python and Tensorflow, software was developed to identify and classify specific cell types in the tissue. In this way, different cell types could be counted automatically for user-defined tissue regions.
Tools:
PythonTensorFlowgit -
Python Developer, DevOps
2018
Development of a parallelized image analysis pipeline for processing massive image data of an automated microscope.
An automation system within a pharmaceutical laboratory produces terabytes of image data daily. Based on CellProfiler software, object recognition and feature calculation was implemented to extract structured data from the raw image data. For robust deployment on an in-house CPU cluster, the application was containerized with Docker and orchestrated with SLURM.
Tasks:
- Definition of the specifications in collaboration with the domain experts
- Planning and acquisition of necessary hardware
- Conception and implementation
- Big Data handling
Tools:
PythonCellProfilerDockerSLURM -
Data Scientist Drug Discovery
2014
Drug Screening Analysis including Feature Engineering, Clustering and Ranking Analysis
Tasks:
- Setup and execution of image analysis pipeline for automated object detection of image based drug screening data.
- Feature processing and selection
- Clustering analyses to identify drug candidate groups
- Development of ranking algorithms to identify drugs with high potential and low toxicity
- Big Data processing
Tools:
PythonCellProfilerPandasNumpyScipyScikitLearnApache Spark