Research map / Detection and segmentation / Weakly supervised detection and segmentation
Weakly supervised object localization
Improve class activation maps and erasing strategies to localize whole objects.
54 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Weakly supervised detection and segmentation, keyword cluster “weakly supervised · supervised semantic · image level”
State History Tried Checklist Preprints Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
7 papers in 2023–24 and 9 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 6.1% to 12.5%.
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 · Weakly Supervised Object Localization With Progressive Domain Adaptation (CVPR 2016 · 225 citations)
- 2017 · WILDCAT: Weakly Supervised Learning of Deep ConvNets for Image Classification, Pointwise Localization and Segmentation (CVPR 2017 · 368 citations)
- 2018 · Adversarial Complementary Learning for Weakly Supervised Object Localization (CVPR 2018 · 603 citations)
- 2019 · Attention-Based Dropout Layer for Weakly Supervised Object Localization (CVPR 2019 · 398 citations)
- 2020 · Evaluating Weakly Supervised Object Localization Methods Right (CVPR 2020 · 191 citations)
- 2021 · Patch2Pix: Epipolar-Guided Pixel-Level Correspondences (CVPR 2021 · 228 citations)
- 2022 · LCTR: On Awakening the Local Continuity of Transformer for Weakly Supervised Object Localization (AAAI 2022 · 55 citations)
- 2023 · LOCATE: Localize and Transfer Object Parts for Weakly Supervised Affordance Grounding (CVPR 2023 · 41 citations)
- 2024 · Learning without Exact Guidance: Updating Large-scale High-resolution Land Cover Maps from Low-resolution Historical Labels (CVPR 2024 · 48 citations)
- 2025 · Reasoning Mamba: Hypergraph-Guided Region Relation Calculating for Weakly Supervised Affordance Grounding (CVPR 2025 · 3 citations)
- 2026 · From Passive Perception to Active Memory: A Weakly Supervised Image Manipulation Localization Framework Driven by Coarse-Grained Annotations (AAAI 2026 · 2 citations)
Earliest papers here
- Weakly Supervised Object Localization With Progressive Domain Adaptation (CVPR 2016 · 225 citations)
- ProNet: Learning to Propose Object-Specific Boxes for Cascaded Neural Networks (CVPR 2016 · 61 citations)
- Cascaded Interactional Targeting Network for Egocentric Video Analysis (CVPR 2016 · 49 citations)
What has been tried
46 papers here report 474 measured changes (a component added, removed or swapped, compared with a baseline): helped 80%, no clear effect 2%, hurt 14%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add the Transformer branch and PLAT to the CNN branch | Chesapeake Bay, six-state average · mIoU (%) | 60.15 -> 64.65 | Helped | Learning without Exact Guidance: Updating Large-scale High-resolution Land Cover Maps from Low-resolution Historical Labels (CVPR 2024 · 48 citations) |
| Add data augmentation during fine-tuning | CUB-200-2011 · MaxBoxAcc | 77.68 -> 74.39; -3.30 | Hurt | CAM Back Again: Large Kernel CNNs from a Weakly Supervised Object Localization Perspective (CVPR 2024 · 9 citations) |
| add geometric concentration loss Lgc | AGD20K-Seen · KLD | 1.732 -> 1.487 | Helped | Reasoning Mamba: Hypergraph-Guided Region Relation Calculating for Weakly Supervised Affordance Grounding (CVPR 2025 · 3 citations) |
| Add label refinement to Ours-baseline | AGD20K seen split · KLD | 0.938 -> 0.916 | Helped | Weakly-Supervised Affordance Grounding Guided by Part-Level Semantic Priors (ICLR 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 5 papers of this approach from 2024–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- AGD20K · 2 of 5 papers
Baselines
- Cross-view-AG · 2 of 5 papers
- LOCATE · 2 of 5 papers
- WSMA · 2 of 5 papers
Recent preprints
Loading recent arXiv preprints…
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
- Weakly supervised semantic segmentation (116 papers)
- Weakly supervised detection (97 papers)
- CAM refinement and prototypes (52 papers)
Elsewhere in the topic
- Pseudo-labeling for semi-supervised detection (76 papers)
- Consistency regularization for segmentation (25 papers)
- Semi-supervised instance segmentation (16 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).