Automatic crack detection in civil infrastructure based on a hybrid fine tuned MnasNet and adaptive patching
Kun Li, Feng He, Xiaojun Zhao · 2025
Read Automatic crack detection in civil infrastructure based on a hybrid fine tuned MnasNet and adaptive patching by Kun Li, Feng He, Xiaojun Zhao (2025) 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-12756241.txt.
Description
Kun Li, Feng He, and Xiaojun Zhao present an automatic crack detector for bridges and other structures in this 2025 paper. Spotting cracks early keeps infrastructure safe and cuts upkeep costs. Older image methods falter in real conditions, and heavy neural models are too slow for field devices.
Their system pairs a light neural network with an optimizer inspired by chaos patterns and long jumps. A smart patching step resizes the image chunks to match scene complexity. On a standard crack dataset, the system topped rival models on accuracy and balance of errors, while checking each image in a twentieth of a second. The authors target phone and embedded inspections.
Catalog record
| CREATOR | Kun Li, Feng He, Xiaojun Zhao |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 85 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |