Bachelor thesis: Digital Pathology Scan Duplicate Classification

03.04.2025, Abschlussarbeiten, Bachelor- und Masterarbeiten

The Computational Pathology Lab of the TUM Institute for Pathology is offering a bachelor’s thesis titled “QC for digital pathology – scan duplicate classification”.

The TUM Institute for Pathology digitizes 200.000 tissue slides per year using high-resolution slide scanner. This process is semi-automatic: physical slides are produced in our lab, loaded into the slides scanners and automatically scanned to whole slide images (WSI). The WSI is then automatically sent to our laboratory information system / database and registered to the right patient via a unique barcode on the slide.

However, errors happen during the manual creation of the slide, e.g. using the same barcode for different slides. Also, a slide can be scanned twice (e.g., due to scanning artifacts) resulting in two WSI with the same barcode. Although duplicates are easily identified at the time of patient sign-out, automated detection of such duplicates is wanted to keep the database clean and facilitate downstream research with clinical data.

This project aims to build an automated duplicate detector based on WSI. The detector will run on our retrospective database differentiating duplicates into rescans of the same slide (one of them can be deleted) or erroneous re-use of the same barcode on different slides (need manual resolution by our lab).

Kontakt: peter.schueffler@tum.de

Mehr Information