Research map / Detection and segmentation / Open-set and few-shot detection
Amodal and open-world instance segmentation
Segment instances, including occluded ones, without annotations using pseudo labels.
137 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Open-set and few-shot detection, keyword cluster “images · anomaly · interactive”
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.
34 papers in 2023–24 and 24 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 5.8% to 4.3%.
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
- 2016 · HD Maps: Fine-Grained Road Segmentation by Parsing Ground and Aerial Images (CVPR 2016 · 150 citations)
- 2017 · Scene Parsing Through ADE20K Dataset (CVPR 2017 · 3,305 citations)
- 2018 · Deep Extreme Cut: From Extreme Points to Object Segmentation (CVPR 2018 · 474 citations)
- 2019 · Panoptic Segmentation (CVPR 2019 · 1,607 citations)
- 2020 · Video Panoptic Segmentation (CVPR 2020 · 169 citations)
- 2021 · Simple Copy-Paste Is a Strong Data Augmentation Method for Instance Segmentation (CVPR 2021 · 1,154 citations)
- 2022 · FreeSOLO: Learning To Segment Objects Without Annotations (CVPR 2022 · 110 citations)
- 2023 · Devil Is in the Queries: Advancing Mask Transformers for Real-World Medical Image Segmentation and Out-of-Distribution Localization (CVPR 2023 · 31 citations)
- 2024 · Unsupervised Universal Image Segmentation (CVPR 2024 · 37 citations)
- 2025 · ROS-SAM: High-Quality Interactive Segmentation for Remote Sensing Moving Object (CVPR 2025 · 25 citations)
- 2026 · Effective Robotic Cloth Grasping Through Suppressing False Discoveries (AAAI 2026 · 5 citations)
Earliest papers here
- HD Maps: Fine-Grained Road Segmentation by Parsing Ground and Aerial Images (CVPR 2016 · 150 citations)
- Active Image Segmentation Propagation (CVPR 2016 · 116 citations)
- Object Co-Segmentation via Graph Optimized-Flexible Manifold Ranking (CVPR 2016 · 85 citations)
What has been tried
129 papers here report 1,358 measured changes (a component added, removed or swapped, compared with a baseline): helped 74%, no clear effect 6%, hurt 18%, 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 |
|---|---|---|---|---|
| Increase cluster count from 800 to 2911 | COCO · APbox50 | 11.8 -> 13.3 | Helped | Unsupervised Universal Image Segmentation (CVPR 2024 · 37 citations) |
| add an occluder predictor | COCOA · mIOU | 69.9 -> 88.4 | Helped | Amodal Ground Truth and Completion in the Wild (CVPR 2024 · 28 citations) |
| Add the proposed inference pipeline to SAM | SAT-MTB · IoU | 37.25 -> 43.41 | Helped | ROS-SAM: High-Quality Interactive Segmentation for Remote Sensing Moving Object (CVPR 2025 · 25 citations) |
| Add depth-only Depth-Guided Reprojection (DGR) to direct BEV projection | OmniCity and Brooklyn, land use and floor · mIOU | 27.07 -> 32.52; 34.13 -> 37.01; 37.04 -> 42.33; 39.61 -> 45.06; 32.10 -> 35.49; 30.56 -> 37.64; 42.43 -> 45.16; 43.55 -> | Helped | SG-BEV: Satellite-Guided BEV Fusion for Cross-View Semantic Segmentation (CVPR 2024 · 25 citations) |
| Replace Area Graph matching with dense area similarity comparison | ScanNet1500 area matching · AOR | 72.75 -> 69.23 | Hurt | MESA: Matching Everything by Segmenting Anything (CVPR 2024 · 23 citations) |
| Scale training data from COCONut-S to COCONut-L | COCONut-val · PQ | 59.4 -> 62.7 | Helped | COCONut: Modernizing COCO Segmentation (CVPR 2024 · 16 citations) |
| Add random fluctuation to anomaly-score values, varying noise from 0% to 4% | SMIYC Anomaly, SMIYC Obstacle, RoadAnomaly · IoU | 0%: 74.57, 66.29, 55.68; 1%: 77.54, 67.64, 58.49; 2%: 83.44, 60.62, 58.67; 3%: 74.89, 56.53, 56.18; 4%: 77.85, 57.73, 56 | Mixed | Segment Every Out-of-Distribution Object (CVPR 2024 · 14 citations) |
| add the feature loss to the semantic decoder | BDDAnomaly, novel class discovery · mIoU [%] | 48.1 -> 62.4; 53.8 -> 62.2; 39.9 -> 56.8 | Helped | Open-World Semantic Segmentation Including Class Similarity (CVPR 2024 · 12 citations) |
| Add SLIC regions to SAM regions | Pascal VOC 2012 val · mIoU | 83.6 -> 87.2 | Helped | Region-Based Representations Revisited (CVPR 2024 · 11 citations) |
| Add Omni2Ego distillation | AuditoryTestPseudo OOV · mIoU (%) | 43.5 -> 46.7 | Helped | Segment beyond View: Handling Partially Missing Modality for Audio-Visual Semantic Segmentation (AAAI 2024 · 9 citations) |
| Add the fusion structure and region branch | KINS · mean IoU | 65.9 -> 73.4 | Helped | BLADE: Box-Level Supervised Amodal Segmentation through Directed Expansion (AAAI 2024 · 8 citations) |
| add quantization-based semantic decomposition | AVS-Object-Multi · J &F | 57.6 -> 59.7 | Helped | QDFormer: Towards Robust Audiovisual Segmentation in Complex Environments with Quantization-based Semantic Decomposition (CVPR 2024 · 8 citations) |
What reviewers will expect
Datasets and baselines named in the 18 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- LVIS · 4 of 18 papers
- COCO · 3 of 18 papers
- SA-1B · 3 of 18 papers
- Cityscapes · 2 of 18 papers
- DAVIS 2017 · 2 of 18 papers
- VOST · 2 of 18 papers
- MOSE · 2 of 18 papers
- YouTube-VOS · 2 of 18 papers
Baselines
- XMem · 3 of 18 papers
- SAM · 3 of 18 papers
- OneFormer · 2 of 18 papers
- SAM2 · 2 of 18 papers
- HQ-SAM · 2 of 18 papers
- SimpleClick · 2 of 18 papers
- AISFormer · 2 of 18 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 58 papers; from OpenAlex and the papers' affiliations).
- Shanghai Jiao Tong University · 5 papers
- Alibaba Group · 5 papers
- The University of Adelaide · 4 papers
- Korea Advanced Institute of Science and Technology · 4 papers
- Carnegie Mellon University · 4 papers
- Adobe · 3 papers
- Zhejiang University · 3 papers
- Sun Yat-sen University · 3 papers
- Mohamed bin Zayed University of Artificial Intelligence · 3 papers
- SenseTime · 2 papers
Use it
Download BibTeX of 40 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
- Object-level representation learning (219 papers)
- Open-set and unknown object detection (160 papers)
- Interactive and referring segmentation (161 papers)
- Industrial anomaly detection (117 papers)
- Remote sensing segmentation and change (84 papers)
Elsewhere in the topic
- Unsupervised object and part discovery (114 papers)
- Domain-adaptive object detection (103 papers)
- Open-vocabulary object detection (96 papers)
- Open-vocabulary semantic segmentation (93 papers)
- Object-centric learning theory and video (86 papers)
- Few-shot object detection (67 papers)
- Human-object interaction detection (66 papers)
- Few-shot segmentation (64 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).