Abstract
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.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Zhang, F., Huang, D., Wen, J., & Xie, L. (2026). QMSR: Query-Conditioned Mask-wise Expert Routing for Robust Open-Vocabulary Underwater Object Retrieval. https://omanscience.com/en/articles/qmsr-query-conditioned-mask-wise-expert-routing-for-robust-open-vocabulary-underwater-object-retrieval
MLA 9
Zhang, Fuming, et al. "QMSR: Query-Conditioned Mask-wise Expert Routing for Robust Open-Vocabulary Underwater Object Retrieval." https://omanscience.com/en/articles/qmsr-query-conditioned-mask-wise-expert-routing-for-robust-open-vocabulary-underwater-object-retrieval.
Chicago (author–date)
Zhang, Fuming, Dongyue Huang, Junjie Wen, and Lihua Xie. 2026. "QMSR: Query-Conditioned Mask-wise Expert Routing for Robust Open-Vocabulary Underwater Object Retrieval." https://omanscience.com/en/articles/qmsr-query-conditioned-mask-wise-expert-routing-for-robust-open-vocabulary-underwater-object-retrieval.
Harvard
Zhang, F., Huang, D., Wen, J. and Xie, L. (2026) 'QMSR: Query-Conditioned Mask-wise Expert Routing for Robust Open-Vocabulary Underwater Object Retrieval', Available at: https://omanscience.com/en/articles/qmsr-query-conditioned-mask-wise-expert-routing-for-robust-open-vocabulary-underwater-object-retrieval.
Vancouver
Zhang F, Huang D, Wen J, Xie L. QMSR: Query-Conditioned Mask-wise Expert Routing for Robust Open-Vocabulary Underwater Object Retrieval. https://omanscience.com/en/articles/qmsr-query-conditioned-mask-wise-expert-routing-for-robust-open-vocabulary-underwater-object-retrieval
IEEE
F. Zhang, D. Huang, J. Wen, and L. Xie, "QMSR: Query-Conditioned Mask-wise Expert Routing for Robust Open-Vocabulary Underwater Object Retrieval," https://omanscience.com/en/articles/qmsr-query-conditioned-mask-wise-expert-routing-for-robust-open-vocabulary-underwater-object-retrieval.