Research map / Detection and segmentation / CNN features
Context for hard detection conditions
Use context and attention to detect camouflaged, low-light, and one-shot objects.
133 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in CNN features, keyword cluster “local · context · semantic segmentation”
State History Benchmarks Tried Baselines Checklist Preprints Who Export
Where it stands
Still moving The best reported result is still improving on 1 of 2 tracked benchmarks.
30 papers in 2023–24 and 38 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 9.1% to 11.0%.
Open questions the data points to
- ImageNet VID validation (mAP): best went from 86.7 (up to 2024) to 90.3 in 2025 — TGBFormer: Transformer-GraphFormer Blender Network for Video Object Detection (AAAI 2025 · 4 citations)
No longer reported by papers here: PASCAL VOC 2007 test (mAP, last in 2019).
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 · Object Contour Detection With a Fully Convolutional Encoder-Decoder Network (CVPR 2016 · 376 citations)
- 2017 · Spindle Net: Person Re-Identification With Human Body Region Guided Feature Decomposition and Fusion (CVPR 2017 · 933 citations)
- 2018 · Towards High Performance Video Object Detection (CVPR 2018 · 302 citations)
- 2019 · Distraction-Aware Shadow Detection (CVPR 2019 · 170 citations)
- 2020 · Memory Enhanced Global-Local Aggregation for Video Object Detection (CVPR 2020 · 359 citations)
- 2021 · LoFTR: Detector-Free Local Feature Matching With Transformers (CVPR 2021 · 1,683 citations)
- 2022 · Target-Aware Dual Adversarial Learning and a Multi-Scenario Multi-Modality Benchmark To Fuse Infrared and Visible for Object Detection (CVPR 2022 · 1,179 citations)
- 2023 · CDDFuse: Correlation-Driven Dual-Branch Feature Decomposition for Multi-Modality Image Fusion (CVPR 2023 · 924 citations)
- 2024 · Poly Kernel Inception Network for Remote Sensing Detection (CVPR 2024 · 528 citations)
- 2025 · Pinwheel-shaped Convolution and Scale-based Dynamic Loss for Infrared Small Target Detection (AAAI 2025 · 199 citations)
- 2026 · MODA: The First Challenging Benchmark for Multispectral Object Detection in Aerial Images (AAAI 2026 · 3 citations)
Earliest papers here
- Object Contour Detection With a Fully Convolutional Encoder-Decoder Network (CVPR 2016 · 376 citations)
- RIFD-CNN: Rotation-Invariant and Fisher Discriminative Convolutional Neural Networks for Object Detection (CVPR 2016 · 175 citations)
- Object Skeleton Extraction in Natural Images by Fusing Scale-Associated Deep Side Outputs (CVPR 2016 · 97 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.
ImageNet VID validation · mAP (higher is better)
4 papers, 2018–2025 · still improving
| Year | Best so far | Set by |
|---|---|---|
| 2018 | 78.6 | Towards High Performance Video Object Detection (CVPR 2018 · 302 citations) |
| 2020 | 82.9 | Memory Enhanced Global-Local Aggregation for Video Object Detection (CVPR 2020 · 359 citations) |
| 2021 | 86.7 | MAMBA: Multi-level Aggregation via Memory Bank for Video Object Detection (AAAI 2021 · 60 citations) |
| 2025 | 90.3 | TGBFormer: Transformer-GraphFormer Blender Network for Video Object Detection (AAAI 2025 · 4 citations) |
PASCAL VOC 2007 test · mAP (higher is better)
4 papers, 2016–2019 · not reported since 2019
| Year | Best so far | Set by |
|---|---|---|
| 2016 | 71 | RIFD-CNN: Rotation-Invariant and Fisher Discriminative Convolutional Neural Networks for Object Detection (CVPR 2016 · 175 citations) |
| 2018 | 76 | Structure Inference Net: Object Detection Using Scene-Level Context and Instance-Level Relationships (CVPR 2018 · 262 citations) |
| 2019 | 82.7 | Feature Intertwiner for Object Detection (ICLR 2019 · 18 citations) |
What has been tried
114 papers here report 1,302 measured changes (a component added, removed or swapped, compared with a baseline): helped 75%, no clear effect 5%, hurt 19%, mixed 1%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| use multi-scale kernels instead of only 3×3 kernels | DOTA-v1.0 · mAP | 76.94 -> 78.16 | Helped | Poly Kernel Inception Network for Remote Sensing Detection (CVPR 2024 · 528 citations) |
| Replace the first two standard convolutions with PConv(4,3) | IRSTD-1K · mAP50 | 87.4 -> 89.9 | Helped | Pinwheel-shaped Convolution and Scale-based Dynamic Loss for Infrared Small Target Detection (AAAI 2025 · 199 citations) |
| remove multi-scale feature encoding (replace PVT with Res2Net-50 and remove ASPP) | ACOD-12K · Sα | 0.866 -> 0.855 | Hurt | Depth-Aware Concealed Crop Detection in Dense Agricultural Scenes (CVPR 2024 · 96 citations) |
| add SOM | DroneVehicle validation · mAP (%) | 76.1 -> 76.5 | Helped | Weakly Misalignment-free Adaptive Feature Alignment for UAVs-based Multimodal Object Detection (CVPR 2024 · 84 citations) |
| Set shunt ratio α to 0 instead of α=1/2 | HRSID test · AP50 | 95.71 % -> 94.80 % | Hurt | Unleashing Channel Potential: Space-Frequency Selection Convolution for SAR Object Detection (CVPR 2024 · 74 citations) |
| Add B-CAS | PASCAL VOC Novel Set 1, 3-shot · AP50 | 56.7 -> 61.2 | Helped | Fine-Grained Prototypes Distillation for Few-Shot Object Detection (AAAI 2024 · 68 citations) |
| Add TPM to the AL baseline | RGBNT201 · mAP | 62.9 -> 67.8 | Helped | TOP-ReID: Multi-Spectral Object Re-identification with Token Permutation (AAAI 2024 · 49 citations) |
| Add HFU | LLVIP · mAP50 | 90.9 -> 92.4 | Helped | FD2-Net: Frequency-Driven Feature Decomposition Network for Infrared-Visible Object Detection (AAAI 2025 · 43 citations) |
| Add adaptive classifier attention | SRD shadow region · RMSE | 6.35 -> 5.88 | Helped | DeS3: Adaptive Attention-Driven Self and Soft Shadow Removal Using ViT Similarity (AAAI 2024 · 38 citations) |
| Add adaptive smoothing to full features without foreground-background decoupling | AI-TOD test · AP | 12.0 -> 12.7 | Helped | SET: Spectral Enhancement for Tiny Object Detection (CVPR 2025 · 28 citations) |
| Add illumination-invariant features with IIM | ExDark · mAP50 | 72.5 -> 74.7 | Helped | You Only Look Around: Learning Illumination-Invariant Feature for Low-light Object Detection (NeurIPS 2024 · 28 citations) |
| Add FPN to low-light image input | DARK FACE · mAP (%) | 61.13 -> 66.42 | Helped | Trash to Treasure: Low-Light Object Detection via Decomposition-and-Aggregation (AAAI 2024 · 27 citations) |
Baseline numbers that disagree
The same baseline on the same benchmark and metric, reported more than 5% apart by different papers here. Training budgets, data splits and reimplementations differ, so state which number you compare with.
What reviewers will expect
Datasets and baselines named in the 25 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- LLVIP · 2 of 25 papers
- GDD · 2 of 25 papers
- GSD · 2 of 25 papers
- VisDrone2019 · 2 of 25 papers
- AI-TOD · 2 of 25 papers
Baselines
- AFSD · 2 of 25 papers
- ActionFormer · 2 of 25 papers
- TADTR · 2 of 25 papers
- Tridet · 2 of 25 papers
- DyFADet · 2 of 25 papers
- Faster R-CNN · 2 of 25 papers
- RFLA · 2 of 25 papers
- SR-TOD · 2 of 25 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 68 papers; from OpenAlex and the papers' affiliations).
- Peng Cheng Laboratory · 4 papers
- Xidian University · 4 papers
- Tsinghua University · 3 papers
- City University of Hong Kong · 3 papers
- Harbin Institute of Technology · 3 papers
- Southern University of Science and Technology · 3 papers
- Nanjing University of Science and Technology · 3 papers
- University of Electronic Science and Technology of China · 3 papers
- ETH Zurich · 3 papers
- Korea University · 2 papers
Use it
Download BibTeX of 39 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
- Context modeling for semantic segmentation (241 papers)
- Local features and context in CNNs (170 papers)
- Domain-specific segmentation and detection (118 papers)
- Forgery detection (46 papers)
- Lane detection (32 papers)
Elsewhere in the topic
- Region-based and query-based detectors (183 papers)
- Convolution-transformer backbones (101 papers)
- Box regression and loss design (94 papers)
- Efficient DETR query design (85 papers)
- Boundary-aware instance segmentation (83 papers)
- Plain and high-resolution ViTs (69 papers)
- Arbitrary-shape scene text detection (61 papers)
- RGB salient object detection (61 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).