A transformer-based multi-modal fusion framework for laryngeal lesion classification using contact endoscopy
Yunjing Duan, Wei Li · 2026
Read A transformer-based multi-modal fusion framework for laryngeal lesion classification using contact endoscopy by Yunjing Duan, Wei Li (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/pmc-13582240.txt.
Description
Yunjing Duan and Wei Li build a framework based on a transformer, to sort laryngeal lesions into classes from contact endoscopy images. Their 2026 study joins data from clinics, radiomic features, and representations from deep learning.
Three hundred patients from three medical centres gave close to eight thousand images. The team joined eighteen clinical variables, more than one hundred radiomic features, and thousands of deep features. They did so through a transformer encoder. More than one way to pick features, and more than one way to sort them into classes, were set side by side. The best set-up used one way to pick features, with a vision transformer classifier. Accuracy stayed high from training, through checks made in the same group, to a test centre that stood on its own. The system aims at sorting of lesions that runs on its own, and that is true.
Catalog record
| CREATOR | Yunjing Duan, Wei Li |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |