Metraipolis › archive › dc_cd9306dbf734
AquaProbe: Efficient Adaptation of Frozen Vision Models for Underwater Images — arXiv:2610.40900
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.40900
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-02T04:52:41.602Z
- SHA-256 of the archived text
c86ea3873f1b011050201d2546e4d716f3c21c2fb62dcadefa441206b6e93183- Stable URL of this page
https://metraipolis.com/documento/dc_cd9306dbf734
How to cite
AquaProbe: Efficient Adaptation of Frozen Vision Models for Underwater Images — arXiv:2610.40900 — arXiv (en), https://arxiv.org/abs/2610.40900 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cd9306dbf734 (Metraipolis archive dc_cd9306dbf734, sha256 c86ea3873f1b011050201d2546e4d716f3c21c2fb62dcadefa441206b6e93183)
BibTeX
@misc{dccd9306dbf734,
title = {AquaProbe: Efficient Adaptation of Frozen Vision Models for Underwater Images — arXiv:2610.40900},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.40900},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cd9306dbf734; sha256 c86ea3873f1b011050201d2546e4d716f3c21c2fb62dcadefa441206b6e93183},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.40900. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reliable visual perception is essential for safe autonomous underwater exploration. Yet vision models pretrained on terrestrial images become less reliable underwater, where turbidity, backscatter, and wavelength-dependent attenuation alter scene appearance. Existing adaptation methods either spread adapters equally across the encoder depth or select layers using the divergence between two fixed domains. In this paper, we introduce AquaProbe for measuring how the location of the domain gap changes with water type and imaging range. AquaProbe separates newly introduced divergence from that accumulated through a frozen vision encoder using paired clean and physically rendered images and allocates a fixed LoRA budget according to the measured profile. The segmentation mean Intersection-over-Union (mIoU) for the proportional LoRA rank allocation, based on the measured profile, is 70.1; this is a better result than uniform LoRA rank allocation by 1.9 mIoU points while having the same parameter budget of 0.59 million adapter parameters. The three-way router, conditioned on estimated condition clusters, achieves 70.6 mIoU on the standard SUIM split and 63.4 AUDRC under held-out degradation. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG); Robotics (cs.RO). Pagina del lavoro: https://arxiv.org/abs/2610.40900 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-02T04:45:12.942Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).