DefectAtlas-20: A real-world image dataset for joint material and surface-condition classification
Ricardo Buettner, Leopold Fischer-Brandies, Lukas Pasold, Luca Eisentraut · 2026
Read DefectAtlas-20: A real-world image dataset for joint material and surface-condition classification by Ricardo Buettner, Leopold Fischer-Brandies, Lukas Pasold, Luca Eisentraut (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-13571519.txt.
Description
Ricardo Buettner and co-authors published this data set paper in 2026. Automated surface checks need photos of damage as it looks in real life, not in a lab. Past sets stuck to one material or staged scenes. DefectAtlas-20 holds 13,233 phone photos in twenty classes: ten materials, each shown sound and damaged. The shots were taken around Hamburg, Germany, indoors and out.
Materials run from bare and coated metal and wood to rubber, cloth, glass, and cable. Light, angle, and backdrop vary as they do in the field. Each photo was labeled by hand, checked by a second eye, and stripped of private data. The set suits tasks from defect spotting to sorting materials. The authors offer it as a bench for inspection tools used in upkeep and repair.
Catalog record
| CREATOR | Ricardo Buettner, Leopold Fischer-Brandies, Lukas Pasold, Luca Eisentraut |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |