A spatial adaptive multi-scale ConvNeXt framework for robust, calibrated, and statistically validated multi-class retinal OCT classification
Mithun Vijayan, Goutham Veerapu, Nisha J. S. · 2026
Read A spatial adaptive multi-scale ConvNeXt framework for robust, calibrated, and statistically validated multi-class retinal OCT classification by Mithun Vijayan, Goutham Veerapu, Nisha J. S. (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-13572616.txt.
Description
Mithun Vijayan, Goutham Veerapu, and Nisha J S present a retinal scan classifier built for trust as well as accuracy in a 2026 paper. Eye imaging models often chase accuracy while ignoring whether their confidence is honest or whether image quality changes break them. Clinicians need both before relying on a tool.
Their framework adds a spatial adaptive multi scale module to a compact network, refining features at several scales. On an eight disease data set it scored high accuracy and agreement scores. Calibration error was measured and corrected by temperature scaling, dropout sampling estimated uncertainty, and robustness held under noise, blur, and lighting shifts.
Catalog record
| CREATOR | Mithun Vijayan, Goutham Veerapu, Nisha J. S. |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 96 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |