An efficient triple-domain YOLOv8 for real-time aluminum profile defect detection
Mohammad Rostami, Afsaneh Fatemi · 2026
Read An efficient triple-domain YOLOv8 for real-time aluminum profile defect detection by Mohammad Rostami, Afsaneh Fatemi (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/PMC13542131.txt.
Description
Mohammad Rostami and Afsaneh Fatemi present, in a 2026 study, a faster object-detection system for spotting surface flaws on aluminium profiles during manufacture. Low contrast, glare, and stretched thin defects defeat standard detectors that read only the ordinary image.
Their design feeds a YOLOv8 detector three views of the same surface: the plain image, a wavelet view that exposes frequency patterns, and a filter view that stresses edges and texture direction. An attention module weighs the three streams for each defect type. On a public benchmark the system scored highest among rivals on both strict and lenient accuracy measures while running above one hundred frames per second, fast enough for production lines.
Catalog record
| CREATOR | Mohammad Rostami, Afsaneh Fatemi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 102510 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |