Research map / Video / Video segmentation and tracking
Mask propagation for unsupervised video segmentation
Segments objects in video with weak, unsupervised or training-free methods using object descriptors and correspondence.
67 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Video segmentation and tracking, keyword cluster “object segmentation · video object · vos”
State History Benchmarks Tried Checklist Preprints Who Export
Where it stands
Quiet None of the tracked benchmarks has a report from 2025–26.
16 papers in 2023–24 and 8 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 3.6% to 1.5%.
No longer reported by papers here: DAVIS 2017 val (J&F, last in 2024); DAVIS 2016 val (J&F, last in 2023); YouTube-VIS 2019 val (AP, last in 2024); YouTube-VIS 2021 val (AP, 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 · Track and Segment: An Iterative Unsupervised Approach for Video Object Proposals (CVPR 2016 · 123 citations)
- 2017 · FusionSeg: Learning to Combine Motion and Appearance for Fully Automatic Segmentation of Generic Objects in Videos (CVPR 2017 · 381 citations)
- 2018 · Efficient Video Object Segmentation via Network Modulation (CVPR 2018 · 389 citations)
- 2019 · See More, Know More: Unsupervised Video Object Segmentation With Co-Attention Siamese Networks (CVPR 2019 · 533 citations)
- 2020 · Learning Video Object Segmentation From Unlabeled Videos (CVPR 2020 · 175 citations)
- 2021 · SSTVOS: Sparse Spatiotemporal Transformers for Video Object Segmentation (CVPR 2021 · 188 citations)
- 2022 · Reliable Propagation-Correction Modulation for Video Object Segmentation (AAAI 2022 · 53 citations)
- 2023 · DropMAE: Masked Autoencoders With Spatial-Attention Dropout for Tracking Tasks (CVPR 2023 · 178 citations)
- 2024 · VideoCutLER: Surprisingly Simple Unsupervised Video Instance Segmentation (CVPR 2024 · 20 citations)
- 2025 · DeNVeR: Deformable Neural Vessel Representations for Unsupervised Video Vessel Segmentation (CVPR 2025 · 7 citations)
- 2026 · Predicting Video Slot Attention Queries from Random Slot-Feature Pairs (AAAI 2026 · 1 citation)
Earliest papers here
- Track and Segment: An Iterative Unsupervised Approach for Video Object Proposals (CVPR 2016 · 123 citations)
- Primary Object Segmentation in Videos via Alternate Convex Optimization of Foreground and Background Distributions (CVPR 2016 · 80 citations)
- Discovering the Physical Parts of an Articulated Object Class From Multiple Videos (CVPR 2016 · 8 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.
DAVIS 2017 val · J&F (higher is better)
8 papers, 2019–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 125.2 | BubbleNets: Learning to Select the Guidance Frame in Video Object Segmentation by Deep Sorting Frames (CVPR 2019 · 53 citations) |
DAVIS 2016 val · J&F (higher is better)
6 papers, 2019–2023 · not reported since 2023
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 159.8 | BubbleNets: Learning to Select the Guidance Frame in Video Object Segmentation by Deep Sorting Frames (CVPR 2019 · 53 citations) |
YouTube-VIS 2019 val · AP (higher is better)
5 papers, 2021–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 44.6 | Video Instance Segmentation using Inter-Frame Communication Transformers (NeurIPS 2021 · 58 citations) |
| 2022 | 47.5 | TubeFormer-DeepLab: Video Mask Transformer (CVPR 2022 · 32 citations) |
| 2023 | 55.3 | Mask-Free Video Instance Segmentation (CVPR 2023 · 25 citations) |
YouTube-VIS 2021 val · AP (higher is better)
5 papers, 2021–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 36.6 | Video Instance Segmentation using Inter-Frame Communication Transformers (NeurIPS 2021 · 58 citations) |
| 2022 | 41.2 | TubeFormer-DeepLab: Video Mask Transformer (CVPR 2022 · 32 citations) |
| 2023 | 41.6 | Mask-Free Video Instance Segmentation (CVPR 2023 · 25 citations) |
What has been tried
58 papers here report 668 measured changes (a component added, removed or swapped, compared with a baseline): helped 78%, no clear effect 5%, hurt 15%, 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 frame size 180 instead of 360 | YouTubeVIS-2019 · AP50 | 50.7 -> 49.9 | Hurt | VideoCutLER: Surprisingly Simple Unsupervised Video Instance Segmentation (CVPR 2024 · 20 citations) |
| Add generalizable Fourier augmentation with SegFormer | DAVIS-2016 · J &F | 85.3 -> 88.2 | Helped | Generalizable Fourier Augmentation for Unsupervised Video Object Segmentation (AAAI 2024 · 11 citations) |
| Add layer separation bootstrapping | XACV test set · Dice | 0.0877 | Helped | DeNVeR: Deformable Neural Vessel Representations for Unsupervised Video Vessel Segmentation (CVPR 2025 · 7 citations) |
| add TCF to the pure transformer baseline | FDST · MAE | 2.45 -> 2.32 | Helped | Efficient Masked AutoEncoder for Video Object Counting and A Large-Scale Benchmark (ICLR 2025 · 0 citations) |
| Add U-Net mask refinement | MOVI-C, MOVI-D, MOVI-E · segmentation performance | Helped | VONet: Unsupervised Video Object Learning With Parallel U-Net Attention and Object-wise Sequential VAE (ICLR 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 6 papers of this approach from 2024–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- YouTubeVIS-2019 · 2 of 6 papers
- YouTubeVIS-2021 · 2 of 6 papers
Baselines
- OCLR · 2 of 6 papers
- MotionGroup · 2 of 6 papers
- CutLER · 2 of 6 papers
- DINO · 2 of 6 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 24 papers; from OpenAlex and the papers' affiliations).
- Zhejiang University · 2 papers
Use it
Download BibTeX of 31 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
- Matching-based video object segmentation (86 papers)
- Video semantic and panoptic segmentation (67 papers)
Elsewhere in the topic
- Self-supervised video representation learning (167 papers)
- Video restoration with temporal priors (159 papers)
- Layered video decomposition and dynamics (140 papers)
- Stochastic video prediction (125 papers)
- Physics-inspired motion prediction (113 papers)
- Sign language and physiological video analysis (99 papers)
- Weakly supervised video anomaly detection (99 papers)
- Spatio-temporal predictive modeling (91 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).