DDNet: disaster damage detection for buildings based on dual-temporal joint attention network
Xiaosan Ge, Lin Zhou, Di Meng · 2025
Read DDNet: disaster damage detection for buildings based on dual-temporal joint attention network by Xiaosan Ge, Lin Zhou, Di Meng (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-12660927.txt.
Description
Xiaosan Ge, Lin Zhou, and Di Meng present DDNet, a network for detecting building damage after disasters. Their 2025 paper uses satellite images taken before and after an event. Fast, accurate damage maps guide emergency response. The task has two parts: finding buildings and grading their damage.
Buildings cover little of each image, so models often miss their edges. That hurts the grading step that follows. DDNet first locates buildings with a network tuned for fine detail and balance. A joint attention module then mines features from both time points. A final step turns pixel results into building-level damage labels. On the xBD dataset, DDNet scored well on both locating and grading.
Catalog record
| CREATOR | Xiaosan Ge, Lin Zhou, Di Meng |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |