发表机构
Nanyang Technological University; The Chinese University of Hong Kong（南洋理工大学; 香港中文大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对水下开放词汇检索中固定增强策略失效的问题，提出QMSR框架，通过查询条件掩码级专家路由自适应选择增强专家并融合原始语义，显著提升检索性能。
AI 中文摘要
开放词汇目标检索在复杂水下环境中仍然具有挑战性。尽管水下图像增强（UIE）可以改善视觉质量，但固定的UIE策略在检索中甚至可能不如原始（Raw）表示，这表明增强不应作为统一的预处理步骤应用。为解决这一问题，我们提出了QMSR，一种查询条件的掩码级专家路由框架，用于水下开放词汇检索。具体而言，QMSR为每个查询-候选对选择一个预训练的UIE专家，并预测连续的原始-专家融合强度，从而在保留有用原始语义的同时实现自适应增强。在训练期间，特权排序预言机提供专家选择和融合强度监督，而退火软路由松弛有助于优化硬Top-1路由策略。实验表明，与图像-查询共享路由器相比，QMSR将NDCG@10提高了17.6%，同时始终优于固定UIE策略，并在保留的查询类别上保持有效。这些结果证明了查询条件和候选特定增强路由对水下开放词汇检索的有效性。
英文摘要
Open-vocabulary object retrieval remains challenging in complex underwater environments. Although underwater image enhancement (UIE) can improve visual quality, fixed UIE strategies may even underperform the Raw representation in retrieval, indicating that enhancement should not be applied as a uniform preprocessing step. To address this problem, we propose \textbf{QMSR}, a query-conditioned mask-wise expert routing framework for underwater open-vocabulary retrieval. Specifically, QMSR selects one pretrained UIE expert for each query--candidate pair and predicts a continuous Raw--Expert fusion strength, enabling adaptive enhancement while preserving useful Raw semantics. During training, a privileged ranking oracle provides expert-selection and fusion-strength supervision, while an annealed soft-routing relaxation facilitates optimization of the hard Top-1 routing policy. Experiments show that QMSR improves NDCG@10 by 17.6\% over an image--query shared router, while consistently outperforming fixed UIE strategies and remaining effective on held-out query categories. These results demonstrate the effectiveness of query-conditioned and candidate-specific enhancement routing for underwater open-vocabulary retrieval.
CommentsThis paper has been submitted to ICRA2027
