Multi-class semantic segmentation of oral cancer tissues on whole-slide images: A SuperPixel enhanced deep learning approach

Fabian León, Anne Champagnac, Mathieu Dupoy, Jean-Michel Fayette, Tamás Dániel Csűry, Laurent Duraffourg, Pierre Saintigny · 2026

Read Multi-class semantic segmentation of oral cancer tissues on whole-slide images: A SuperPixel enhanced deep learning approach by Fabian León, Anne Champagnac, Mathieu Dupoy, Jean-Michel Fayette, Tamás Dániel Csűry, Laurent Duraffourg, Pierre Saintigny (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13524623.txt.

Description

León and colleagues presented in 2026 a computer method that labels ten tissue types in microscope slides of oral cancer. Pathologists now mark these slides by hand, which is slow, and rare tissue types get few examples to learn from. The method aims to cut that hand work while keeping the labels sharp at tissue borders.

The team grouped nearby pixels into small clusters called superpixels, and used them both to build training labels and to clean up the final maps. They trained on slides from ten patients and tested on slides from two more. Across nine model setups, their best runs reached mean accuracy near 92 percent, with strong overlap scores even for rare tissue classes. The tool plugs into the free QuPath program, so labs can rebuild whole-slide maps with it.

Catalog record

CREATORFabian León, Anne Champagnac, Mathieu Dupoy, Jean-Michel Fayette, Tamás Dániel Csűry, Laurent Duraffourg, Pierre Saintigny
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
LICENSEhttps://creativecommons.org/licenses/by/4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/article/MED/PMC13524623

Browse the full library · About the Tower of Babel