Unmixing microinfrared spectroscopic images of cross-sections of historical oil paintings
Shivam Pande 1 , Nicolas Nadisic 1 2 , Francisco Mederos-Henry 2 3 and Aleksandra Pižurica 1 · 2026
Read Unmixing microinfrared spectroscopic images of cross-sections of historical oil paintings by Shivam Pande 1 , Nicolas Nadisic 1 2 , Francisco Mederos-Henry 2 3 and Aleksandra Pižurica 1 (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-2603-06673.txt.
Description
Pande and team address imaging of artefacts, where scope records a spectrum per pixel, but mixed species in worn samples make manual library matching slow. They propose a self-taught autoencoder for blind splitting, gauging endmember spectra and abundance maps while using spatial structure through patch-based model. A weighted spectral angle distance loss with weights from spatial flatness, neighbour agreement, and roughness reduces feeling to flaws across more than 1500 bands. Next to angle training, it improves clarity in contamination-prone regions, shown on a Ghent Altarpiece section.
Catalog record
| CREATOR | Shivam Pande 1 , Nicolas Nadisic 1 2 , Francisco Mederos-Henry 2 3 and Aleksandra Pižurica 1 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22668 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.06673