UW-VOS: A Large-Scale Dataset for Underwater Video Object Segmentation

TL;DR

This paper introduces UW-VOS, a large-scale underwater video object segmentation dataset, and proposes SAM-U, a lightweight domain-adaptive framework that significantly improves segmentation performance in underwater environments.

Contribution

The paper presents the first large-scale underwater VOS dataset and a novel parameter-efficient adaptation framework for underwater video segmentation.

Findings

UW-VOS contains 1,431 videos and 309,295 annotations across 409 categories.

Existing methods drop in performance by 13 points on UW-VOS, highlighting domain challenges.

SAM-U achieves state-of-the-art results with only 2% trainable parameters.

Abstract

Underwater Video Object Segmentation (VOS) is essential for marine exploration, yet open-air methods suffer significant degradation due to color distortion, low contrast, and prevalent camouflage. A primary hurdle is the lack of high-quality training data. To bridge this gap, we introduce $\textbf{UW-VOS}$, the first large-scale underwater VOS benchmark comprising 1,431 video sequences across 409 categories with 309,295 mask annotations, constructed via a semi-automatic data engine with rigorous human verification. We further propose $\textbf{SAM-U}$, a parameter-efficient framework that adapts SAM2 to the underwater domain. By inserting lightweight adapters into the image encoder, SAM-U achieves state-of-the-art performance with only $\sim$2$\%$ trainable parameters. Extensive experiments reveal that existing methods experience an average 13-point $\mathcal{J}\&\mathcal{F}$ drop on UW-VOS, while SAM-U effectively bridges this domain gap. Detailed attribute-based analysis further identifies small targets, camouflage, and exit-re-entry as critical bottlenecks, providing a roadmap for future research in robust underwater perception.