Research map / Detection and segmentation / Open-set and few-shot detection
Prompted SAM segmentation
Adapt Segment Anything for semantic, point-prompted, and efficient few-shot segmentation.
39 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Open-set and few-shot detection, keyword cluster “zero · shot object · shot segmentation”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
18 papers in 2023–24 and 16 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 3.1% to 2.9%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2017 · AGA: Attribute-Guided Augmentation (CVPR 2017 · 115 citations)
- 2020 · Symmetry and Group in Attribute-Object Compositions (CVPR 2020 · 116 citations)
- 2021 · Learning To Count Everything (CVPR 2021 · 172 citations)
- 2022 · Image Segmentation Using Text and Image Prompts (CVPR 2022 · 527 citations)
- 2023 · Segment Anything in High Quality (NeurIPS 2023 · 109 citations)
- 2024 · EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anything (CVPR 2024 · 240 citations)
- 2025 · Fields of The World: A Machine Learning Benchmark Dataset for Global Agricultural Field Boundary Segmentation (AAAI 2025 · 16 citations)
- 2026 · Tuning-Free Amodal Segmentation via the Occlusion-Free Bias of Inpainting Models (AAAI 2026 · 1 citation)
Earliest papers here
- AGA: Attribute-Guided Augmentation (CVPR 2017 · 115 citations)
- Symmetry and Group in Attribute-Object Compositions (CVPR 2020 · 116 citations)
- Learning To Count Everything (CVPR 2021 · 172 citations)
What has been tried
28 papers here report 330 measured changes (a component added, removed or swapped, compared with a baseline): helped 78%, no clear effect 4%, hurt 16%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| replace MSE reconstruction loss with cosine similarity loss | ImageNet-1K · top-1 accuracy | 76.8 -> 76.1 | Hurt | EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anything (CVPR 2024 · 240 citations) |
| Add appearance and geometry matching scores to semantic matching | BOP seven core datasets (mean) · mAP | 44.0 -> 48.1 | Helped | SAM-6D: Segment Anything Model Meets Zero-Shot 6D Object Pose Estimation (CVPR 2024 · 123 citations) |
| add contrastive prototype learning rather than using fixed class prototypes | EndoVis2018, n=2 · Challenge IoU | 76.38 -> 80.33 | Helped | SurgicalSAM: Efficient Class Promptable Surgical Instrument Segmentation (AAAI 2024 · 107 citations) |
| Apply scale-aware fine-tuning with two learnable mask weights | PerSeg · mIoU | +2.7% | Helped | Personalize Segment Anything Model with One Shot (ICLR 2024 · 65 citations) |
| use only point queries instead of both generic and point queries | COCO · PQ | 49.6 -> 45.2 | Hurt | Visual In-Context Prompting (CVPR 2024 · 31 citations) |
| train end-to-end instead of using two-stage training | COCO · AP | 30.2 -> 31.2 | Helped | Semantic-aware SAM for Point-Prompted Instance Segmentation (CVPR 2024 · 26 citations) |
| add global-local feature fusion | REC-8K val · MAE | 7.31 -> 6.80 | Helped | Referring Expression Counting (CVPR 2024 · 20 citations) |
| Use 3-class masks rather than 2-class masks | Slovenia, France, South Africa · Object recall | SVN 0.06 -> 0.07; FRA 0.29 -> 0.45; ZAF 0.24 -> 0.35 | Helped | Fields of The World: A Machine Learning Benchmark Dataset for Global Agricultural Field Boundary Segmentation (AAAI 2025 · 16 citations) |
| Use dynamic MoE expert selection instead of simultaneous multi-scale feature fusion | ISIC 2017 · Jac | 77.4 -> 77.9 | Helped | Convolution Meets LoRA: Parameter Efficient Finetuning for Segment Anything Model (ICLR 2024 · 12 citations) |
| Add Ray-Pair RGB loss | Mip-NeRF 360 · mIoU | 91.0 -> 91.3 | Helped | SANeRF-HQ: Segment Anything for NeRF in High Quality (CVPR 2024 · 11 citations) |
| Omit S+max from point selection | COCO-20i 1-shot · mIoU | 58.7 -> 53.1 | Hurt | Bridge the Points: Graph-based Few-shot Segment Anything Semantically (NeurIPS 2024 · 10 citations) |
| Add SPI to the baseline | FSS-1000 1-shot · mean-IoU | 73.8 -> 77.9 | Helped | SAM-Aware Graph Prompt Reasoning Network for Cross-Domain Few-Shot Segmentation (AAAI 2025 · 7 citations) |
What reviewers will expect
Datasets and baselines named in the 10 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- SAM · 2 of 10 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 34 papers; from OpenAlex and the papers' affiliations).
- Microsoft · 2 papers
- ETH Zurich · 2 papers
- Meta · 2 papers
- Simon Fraser University · 2 papers
- University of Chinese Academy of Sciences · 2 papers
- Google · 2 papers
- Singapore University of Technology and Design · 2 papers
- The Chinese University of Hong Kong · 2 papers
- Carnegie Mellon University · 2 papers
- Hong Kong University of Science and Technology · 2 papers
Use it
Download BibTeX of 28 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Few-shot object detection (67 papers)
- Few-shot segmentation (64 papers)
- Zero-shot segmentation and detection (49 papers)
Elsewhere in the topic
- Object-level representation learning (219 papers)
- Interactive and referring segmentation (161 papers)
- Open-set and unknown object detection (160 papers)
- Amodal and open-world instance segmentation (137 papers)
- Industrial anomaly detection (117 papers)
- Unsupervised object and part discovery (114 papers)
- Domain-adaptive object detection (103 papers)
- Open-vocabulary object detection (96 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).