Temporally-Aware Supervised Contrastive Learning for Polyp Counting in Colonoscopy
Luca Parolari✉, Andrea Cherubini, Lamberto Ballan, Carlo Biffi · 2025
Read Temporally-Aware Supervised Contrastive Learning for Polyp Counting in Colonoscopy by Luca Parolari✉, Andrea Cherubini, Lamberto Ballan, Carlo Biffi (2025) 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/arxiv-2507-02493.txt.
Description
Four researchers in Italy present a better way to count growths found during colon exams, in a 2025 paper. During a colonoscopy, a camera moves through the bowel and may spot polyps, small growths that can turn harmful. Counting them fairly matters for reports and quality checks, but the same growth can appear in many frames. Programs must group frames that show the same growth without merging two look-alike growths.
Older methods leaned on looks alone and ignored when each frame was taken. The new method trains its features with time in mind, so frames of one growth taken close together pull together. A time rule in the grouping step also blocks false links between look-alike growths seen far apart. On public data sets the method cut split counts to less than half the rate of prior work, setting a new best.
Catalog record
| CREATOR | Luca Parolari✉, Andrea Cherubini, Lamberto Ballan, Carlo Biffi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-02493