Research map / Detection and segmentation / Open-set and few-shot detection
Unsupervised object and part discovery
Discover objects and parts without labels through contrastive reconstruction and background cues.
114 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Open-set and few-shot detection, keyword cluster “human · interaction · hoi”
State History Benchmarks Tried Checklist Preprints Who Export
Where it stands
Few recent reports The tracked benchmarks have at most one report from 2025–26 that could beat the old record, too few to say whether they are still moving.
28 papers in 2023–24 and 9 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 4.8% to 1.6%.
No longer reported by papers here: DUT-OMRON (IoU, last in 2024); ECSSD (IoU, last in 2024).
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 · Learning Compact Binary Descriptors With Unsupervised Deep Neural Networks (CVPR 2016 · 382 citations)
- 2017 · Learning Features by Watching Objects Move (CVPR 2017 · 491 citations)
- 2018 · DeblurGAN: Blind Motion Deblurring Using Conditional Adversarial Networks (CVPR 2018 · 1,784 citations)
- 2019 · Learning Unsupervised Video Object Segmentation Through Visual Attention (CVPR 2019 · 254 citations)
- 2020 · Unsupervised Deep Learning via Affinity Diffusion (AAAI 2020 · 15 citations)
- 2021 · VirTex: Learning Visual Representations From Textual Annotations (CVPR 2021 · 355 citations)
- 2022 · Deep Spectral Methods: A Surprisingly Strong Baseline for Unsupervised Semantic Segmentation and Localization (CVPR 2022 · 145 citations)
- 2023 · Cut and Learn for Unsupervised Object Detection and Instance Segmentation (CVPR 2023 · 185 citations)
- 2024 · Guided Slot Attention for Unsupervised Video Object Segmentation (CVPR 2024 · 25 citations)
- 2025 · MOS: Modeling Object-Scene Associations in Generalized Category Discovery (CVPR 2025 · 5 citations)
- 2026 · VGGT-Segmentor: Geometry-Enhanced Cross-View Segmentation (CVPR 2026 · 0 citations)
Earliest papers here
- Learning Compact Binary Descriptors With Unsupervised Deep Neural Networks (CVPR 2016 · 382 citations)
- Optical Flow With Semantic Segmentation and Localized Layers (CVPR 2016 · 148 citations)
- Soft-Segmentation Guided Object Motion Deblurring (CVPR 2016 · 108 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
Cityscapes · mIoU (higher is better)
4 papers, 2021–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 78.8 | What Makes Instance Discrimination Good for Transfer Learning? (ICLR 2021 · 82 citations) |
DUT-OMRON · IoU (higher is better)
4 papers, 2021–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 0.464 | Object Segmentation Without Labels with Large-Scale Generative Models (ICML 2021 · 6 citations) |
| 2022 | 0.615 | MOVE: Unsupervised Movable Object Segmentation and Detection (NeurIPS 2022 · 6 citations) |
| 2024 | 0.646 | HEAP: Unsupervised Object Discovery and Localization with Contrastive Grouping (AAAI 2024 · 3 citations) |
ECSSD · IoU (higher is better)
4 papers, 2021–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 0.684 | Object Segmentation Without Labels with Large-Scale Generative Models (ICML 2021 · 6 citations) |
| 2022 | 0.83 | MOVE: Unsupervised Movable Object Segmentation and Detection (NeurIPS 2022 · 6 citations) |
What has been tried
107 papers here report 1,146 measured changes (a component added, removed or swapped, compared with a baseline): helped 68%, no clear effect 7%, hurt 23%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use guided slots instead of randomly initialized slots | DAVIS-16 · GM | 83.7 -> 84.1 | Helped | Guided Slot Attention for Unsupervised Video Object Segmentation (CVPR 2024 · 25 citations) |
| Add viewpoint regularization Lvp | SPair-71k · PCK@0.1 | 58.6 -> 63.6 | Helped | Improving Semantic Correspondence with Viewpoint-Guided Spherical Maps (CVPR 2024 · 21 citations) |
| Add SAM encoder in place of FlowFormer encoder | Sintel(train) · clean EPE | 0.94 -> 0.89 | Helped | SAMFlow: Eliminating Any Fragmentation in Optical Flow with Segment Anything Model (AAAI 2024 · 18 citations) |
| add learnable Reference logits | PASCAL VOC 2012 · mIoU | 49.5 -> 56.7 | Helped | Learn to Rectify the Bias of CLIP for Unsupervised Semantic Segmentation (CVPR 2024 · 12 citations) |
| Add semantic augmentation | KITTI-2015 test · Fl-all (%) | 9.60 -> 9.05 | Helped | UnSAMFlow: Unsupervised Optical Flow Guided by Segment Anything Model (CVPR 2024 · 11 citations) |
| Add 3D-LHP to Depth-Feature Correlation + FPS | COCO-Stuff 27 · unsupervised mIoU | 25.6 -> 26.7 | Helped | Unsupervised Semantic Segmentation Through Depth-Guided Feature Correlation and Sampling (CVPR 2024 · 8 citations) |
| replace MaskCut proposals with VoteCut proposals | COCO val2017 · APmask50 | 15.8 -> 17.7 | Helped | CuVLER: Enhanced Unsupervised Object Discoveries through Exhaustive Self-Supervised Transformers (CVPR 2024 · 7 citations) |
| Add translation decoupling | PartNet · Chamfer Distance | 0.590 -> 0.506 | Helped | ShapeMatcher: Self-Supervised Joint Shape Canonicalization Segmentation Retrieval and Deformation (CVPR 2024 · 7 citations) |
| use object images for traditional training | CUB All · ACC | 61.50 -> 63.06 | Helped | MOS: Modeling Object-Scene Associations in Generalized Category Discovery (CVPR 2025 · 5 citations) |
| add Lfs to Lsmo | Cityscapes · Acc. | 75.7 -> 81.1 | Helped | Integrating Low-Level Visual Cues for Enhanced Unsupervised Semantic Segmentation (AAAI 2025 · 4 citations) |
| Add cluster consistency loss | nuScenes mini, dynamic points · EPE | 0.201 -> 0.085 | Helped | Floxels: Fast Unsupervised Voxel Based Scene Flow Estimation (CVPR 2025 · 4 citations) |
| Add trajectory loss to flow loss | DAVIS · J | 78.5 -> 81.7 | Helped | Learning Segmentation from Point Trajectories (NeurIPS 2024 · 4 citations) |
What reviewers will expect
Datasets and baselines named in the 7 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- COCO-Stuff · 2 of 7 papers
- ImageNet-100 · 2 of 7 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 37 papers; from OpenAlex and the papers' affiliations).
- Peng Cheng Laboratory · 3 papers
- Meta · 3 papers
- Tsinghua University · 2 papers
- Peking University · 2 papers
- University of Michigan · 2 papers
- EPFL · 2 papers
- University of Oxford · 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
- Human-object interaction detection (66 papers)
- Visual relationship and scene graph detection (63 papers)
- Hand-object and articulated interaction (52 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)
- Domain-adaptive object detection (103 papers)
- Open-vocabulary object detection (96 papers)
- Open-vocabulary semantic segmentation (93 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).