Pretext Matters: An Empirical Study of SSL Methods in Medical Imaging
Vedrana Ivezić *, Mara Pleasure *, Ashwath Radhachandran, Saarang Panchavati, Shreeram Athreya, Vivek Sant, Benjamin Emert, Gregory Fishbein, Corey Arnold, William Speier · 2026
Read Pretext Matters: An Empirical Study of SSL Methods in Medical Imaging by Vedrana Ivezić *, Mara Pleasure *, Ashwath Radhachandran, Saarang Panchavati, Shreeram Athreya, Vivek Sant, Benjamin Emert, Gregory Fishbein, Corey Arnold, William Speier (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/arxiv-2603-22649.txt.
Description
Ivezić and co-authors compare self-supervised learning methods for medical images. Such methods learn from unlabeled data, but the best choice varies by domain. The team tests joint embedding methods against predictive variants on ultrasound and histopathology images. Where key signal sits in a small area, view-invariant methods work best. Where signal spans the whole image, as in liver ultrasound, predictive methods win. Expert clinicians checked the value of the learned features. The authors offer rules for matching method to image type.
Catalog record
| CREATOR | Vedrana Ivezić *, Mara Pleasure *, Ashwath Radhachandran, Saarang Panchavati, Shreeram Athreya, Vivek Sant, Benjamin Emert, Gregory Fishbein, Corey Arnold, William Speier |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52307 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.22649