Class-incremental few-shot underwater object detection framework
Xue Zhao, Bin Zhou, Yanjiang Wang · 2026
Read Class-incremental few-shot underwater object detection framework by Xue Zhao, Bin Zhou, Yanjiang Wang (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-13369365.txt.
Description
Xue Zhao, Bin Zhou, and Yanjiang Wang propose a framework for spotting underwater objects when new classes arrive with only a few examples, in a 2026 paper. Underwater detection serves marine exploration and ecological monitoring. It faces three linked problems: poor image quality, shifting species mixes, and scarce labeled samples. Past systems struggle to handle all three at once, which is the gap the framework targets.
Their design joins an image enhancement module with the detector in one end-to-end system, sharpening features from murky input. A prototype-based module with attention at several scales strengthens the class prototypes. A dynamic, boundary-aware scheme then helps prototypes for old and new classes work together. The goal is steady learning as new species appear, without retraining from scratch.
Catalog record
| CREATOR | Xue Zhao, Bin Zhou, Yanjiang Wang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 159 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |