Research map / Detection and segmentation / Open-set and few-shot detection
MarineInst: A Foundation Model for Marine Image Analysis with Instance Visual Description
Ziqiang Zheng, Yiwei Chen, Huimin Zeng, Tuan-Anh Vu, Binh-Son Hua, Sai-Kit Yeung
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add binary instance filtering in the mask decoder (vs fine-tuning SAM only) (ablation) | MarineInst · Marine instance segmentation (automatic) | AP segm: 24.8 -> 30.1 | Improved |
| add model-generated instance masks to training (ablation) | MarineInst · Marine instance segmentation (automatic) | AP segm: 30.1 -> 32.7 | Improved |
| add model-generated instance masks to training (ablation) | SAM-F · Marine instance segmentation (automatic) | AP segm: 24.8 -> 25.8 | Improved |
| continue pre-training SAM on MarineInst20M with human masks only (no filtering) (ablation) | SAM · Marine instance segmentation (automatic) | AP segm: 5.8 -> 24.8 | Improved |
| use MarineInst instead of SAM with point prompts | SAM · Marine instance segmentation (point prompt) | AP segm: 63.0 -> 75.4 | Improved |
| use MarineInst instead of SAM with box prompts | SAM · Marine instance segmentation (box prompt) | AP segm: 93.5 -> 95.4 | Improved |
| use MarineInst-pretrained encoder instead of SAM for salient object segmentation | SAM · USOD10K (ViT-H) | Sm: 0.9034 -> 0.9103 | Improved |
| use marine-specific MarineGPT instead of BLIP2 for instance captions (ablation) | MarineInst captioning · Caption user study | satisfactory score (1-5): 2.31 -> 3.52 | Improved |
| use mask ID prompts in crowded scenes | MarineInst instruction following · instruction-following instance understanding | qualitative | Hurt |
Closest papers by meaning
- SeafloorAI: A Large-scale Vision-Language Dataset for Seafloor Geological Survey (NeurIPS 2024 · similarity 0.86)
- Enhancing Vision-Language Models with Morphological and Taxonomic Knowledge: Towards Coral Recognition for Ocean Health (AAAI 2025 · similarity 0.85)
- CoralSCOP: Segment any COral Image on this Planet (CVPR 2024 · similarity 0.85)
- ReefNet: A Large-Scale Dataset and Benchmark for Fine-Grained Coral Reef Recognition (NeurIPS 2026 · similarity 0.84)
- Fantastic Animals and Where to Find Them: Segment Any Marine Animal with Dual SAM (CVPR 2024 · similarity 0.84)
- NAUTILUS: A Large Multimodal Model for Underwater Scene Understanding (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
