Distinction of cell populations in cervix samples (PAP smears) by using digital holographic microscopy (DHM)
Project and Thesis Description
Cervical cancer is the second most common cancer in women worldwide. In Germany, the standard annual screening procedure is a PAP smear, in which cells are taken by brush biopsy. The current gold standard is a laborious and subjective process involving fixation, staining and examination by a pathologist. We are working on a streamlined and rapid alternative using digital holographic microscopy. This technology provides morphological features and 3D cell data. To derive diagnostically relevant information, it is necessary to distinguish and quantify the different cell populations within a sample, including erythrocytes, leukocytes, epithelial and endothelial cells.
Task Description
- Enhance the current image processing pipeline and introduce new descriptive features.
- Train and validate classical and modern machine learning classifiers to distinguish erythrocytes, leukocytes, epithelial and endothelial cells.
- Explore and adapt non-linear embedding methods to achieve meaningful visualization and clustering of cell populations.
- Include your results (algorithms, visualizations, ...) in the central Python library of the project for sustainable use.
Requirements
- Intermediate knowledge of Python and machine learning libraries
- Interest in healthcare
- Intermediate knowledge of version control tools (git)
- High motivation and independent, entrepreneurial work style, proactive and self-motivated commitment
- Team player and passion for research
Our offerings
- Integration into an inspiring team of interdisciplinary researchers, engineers and entrepreneurs.
- A unique environment for the development of new IVD products with access to all the benefits of the clinical, biomedical, and engineering domains at TranslaTUM.
- Unlimited access to the latest information in the biotechnology industry, as you will work with several stakeholders who are all experts in their field.
Contact CellFace
cellface@ei.tum.de
wiki.tum.de/display/cellface/
Chair of Biomedical Electronics
Ellen Emken
ellen.emken@tum.de