Research map / Detection and segmentation / Open-set and few-shot detection
Remote sensing segmentation and change
Build datasets and models for geospatial segmentation and change detection.
84 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.
26 papers in 2023–24 and 39 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 4.4% to 7.0%.
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 · Cataloging Public Objects Using Aerial and Street-Level Images - Urban Trees (CVPR 2016 · 183 citations)
- 2017 · Fine-Grained Car Detection for Visual Census Estimation (AAAI 2017 · 81 citations)
- 2018 · The INaturalist Species Classification and Detection Dataset (CVPR 2018 · 1,482 citations)
- 2019 · IP102: A Large-Scale Benchmark Dataset for Insect Pest Recognition (CVPR 2019 · 541 citations)
- 2020 · Foreground-Aware Relation Network for Geospatial Object Segmentation in High Spatial Resolution Remote Sensing Imagery (CVPR 2020 · 342 citations)
- 2021 · PPR10K: A Large-Scale Portrait Photo Retouching Dataset With Human-Region Mask and Group-Level Consistency (CVPR 2021 · 55 citations)
- 2022 · ZeroWaste Dataset: Towards Deformable Object Segmentation in Cluttered Scenes (CVPR 2022 · 73 citations)
- 2023 · SAMRS: Scaling-up Remote Sensing Segmentation Dataset with Segment Anything Model (NeurIPS 2023 · 49 citations)
- 2024 · Rotated Multi-Scale Interaction Network for Referring Remote Sensing Image Segmentation (CVPR 2024 · 92 citations)
- 2025 · FoMo: Multi-Modal, Multi-Scale and Multi-Task Remote Sensing Foundation Models for Forest Monitoring (AAAI 2025 · 13 citations)
- 2026 · RS2-SAM2: Customized SAM2 for Referring Remote Sensing Image Segmentation (AAAI 2026 · 4 citations)
Earliest papers here
- Cataloging Public Objects Using Aerial and Street-Level Images - Urban Trees (CVPR 2016 · 183 citations)
- From Tweets to Wellness: Wellness Event Detection from Twitter Streams (AAAI 2016 · 64 citations)
- Fine-Grained Car Detection for Visual Census Estimation (AAAI 2017 · 81 citations)
What has been tried
60 papers here report 492 measured changes (a component added, removed or swapped, compared with a baseline): helped 71%, no clear effect 5%, hurt 19%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add IIM while omitting CIM | RRSIS-D val · mIoU | 61.46 -> 62.27 | Helped | Rotated Multi-Scale Interaction Network for Referring Remote Sensing Image Segmentation (CVPR 2024 · 92 citations) |
| add box prompt | Beijing · IoU | 0.708 -> 0.721 | Helped | UV-SAM: Adapting Segment Anything Model for Urban Village Identification (AAAI 2024 · 57 citations) |
| Add negative-mask training to CoralSCOP | Coral mask generation, automatic · IoU | 38.15 -> 46.46 | Helped | CoralSCOP: Segment any COral Image on this Planet (CVPR 2024 · 38 citations) |
| Use one shared linear projection rather than separate projections for each spectral band | BigEarthNet · F1 micro | 66.11 (±0.07) -> 73.42 (±0.10) | Helped | FoMo: Multi-Modal, Multi-Scale and Multi-Task Remote Sensing Foundation Models for Forest Monitoring (AAAI 2025 · 13 citations) |
| add HFG and FR to the model | Mara-Wildlife · mAP | 0.46 -> 0.54 | Helped | WildlifeMapper: Aerial Image Analysis for Multi-Species Detection and Identification (CVPR 2024 · 11 citations) |
| Add coarse-grained pseudo labels to Source Only | Wildfires (California, USA) · F1-score | 0.7376 -> 0.7396 | Helped | Generalizable Disaster Damage Assessment via Change Detection with Vision Foundation Model (AAAI 2025 · 9 citations) |
| Add the proposed initialization scheme instead of fully random initialization | Open-Canopy test · MAE | 11.17 -> 2.52 | Helped | Open-Canopy: Towards Very High Resolution Forest Monitoring (CVPR 2025 · 8 citations) |
| Add pancreas subregions and 24 surrounding anatomical structures to tumor and pancreas labels | PanTS test set · mean DSC | 57.4% -> 67.7% (+10.3%) | Helped | PanTS: The Pancreatic Tumor Segmentation Dataset (NeurIPS 2025 · 7 citations) |
| add the alignment network to baseline DiR | Set14 · PSNR | 22.4 -> 23.9 | Helped | Learning Degradation-Independent Representations for Camera ISP Pipelines (CVPR 2024 · 7 citations) |
| Use a larger SwinT-L backbone instead of smaller backbones | STVchrono sequential instance segmentation · AP | No clear effect | The STVchrono Dataset: Towards Continuous Change Recognition in Time (CVPR 2024 · 7 citations) | |
| Add unlabeled ground images during the second training phase | MAVREC aerial test · AP | 23.2 -> 26.7 | Helped | Multiview Aerial Visual RECognition (MAVREC): Can Multi-view Improve Aerial Visual Perception? (CVPR 2024 · 6 citations) |
| Add ImageNet1K_V2 pretraining to ResNet-50 | HarvestNet hand-labeled test set · accuracy | 65.05 -> 79.18 | Helped | HarvestNet: A Dataset for Detecting Smallholder Farming Activity Using Harvest Piles and Remote Sensing (AAAI 2024 · 5 citations) |
What reviewers will expect
Datasets and baselines named in the 26 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- DeepLabv3 · 8 of 26 papers
- UNet · 7 of 26 papers
- Faster R-CNN · 4 of 26 papers
- RetinaNet · 3 of 26 papers
- PSPNet · 3 of 26 papers
- SegFormer · 3 of 26 papers
- ViT · 2 of 26 papers
- ConvNext · 2 of 26 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 65 papers; from OpenAlex and the papers' affiliations).
- McGill University · 4 papers
- Hong Kong University of Science and Technology · 4 papers
- Tsinghua University · 4 papers
- The University of Hong Kong · 3 papers
- Sun Yat-sen University · 3 papers
- Wuhan University · 2 papers
- Stanford University · 2 papers
- University of Science and Technology of China · 2 papers
- Indian Institute of Technology · 2 papers
- Mila · 2 papers
Use it
Download BibTeX of 37 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)
- Amodal and open-world instance segmentation (137 papers)
- Industrial anomaly detection (117 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).