Research map / Detection and segmentation
CNN features: research map
1,734 accepted papers on CNN features in Detection and segmentation, from ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026), grouped into 6 clusters and 21 approaches. The busiest year so far is 2022.
Within Detection and segmentation, its share grew from 26.1% in 2023–24 to 28.8% in 2025–26 (328 → 347 papers at ICML, NeurIPS, CVPR and AAAI, the venues with data for all four years).
Explore CNN features in the interactive map
Working on something in this topic? Describe your idea in scime atlas to see which approach it falls under, the closest papers by meaning and how crowded the spot has become.
Approaches and key papers
local · context · semantic segmentation · 740 papers
Approaches in this cluster:
- Context modeling for semantic segmentation (241 papers)
Capture context with structured models, prototypes, and CRFs for semantic segmentation. - Local features and context in CNNs (170 papers)
Learn local descriptors and contextual relations to aid detection and recognition. - Context for hard detection conditions (133 papers)
Use context and attention to detect camouflaged, low-light, and one-shot objects. - Domain-specific segmentation and detection (118 papers)
Adapt context-aware networks to radar, maps, HDR, and medical imaging. - Forgery detection (46 papers)
Detect manipulated faces and images with frequency, noise, and local anomaly cues. - Lane detection (32 papers)
Detect lanes with transformers, global-local perception, and sequence generation.
Most cited and most cited since 2024:
- Pyramid Scene Parsing Network (CVPR 2017 · 15,903 citations)
- Learning Deep Features for Discriminative Localization (CVPR 2016 · 11,042 citations)
- Poly Kernel Inception Network for Remote Sensing Detection (CVPR 2024 · 528 citations)
- Rethinking the Up-Sampling Operations in CNN-based Generative Network for Generalizable Deepfake Detection (CVPR 2024 · 200 citations)
detectors · detr · search · 410 papers
Approaches in this cluster:
- Region-based and query-based detectors (183 papers)
Design detectors using region proposals, learnable queries, and transformer pretraining. - Box regression and loss design (94 papers)
Improve localization with refined losses and anchor-free or rotated box formulations. - Efficient DETR query design (85 papers)
Speed up and ease query competition in DETR-style detectors. - Pedestrian and crowd detection (48 papers)
Detect people in crowds with specialized losses and datasets.
Most cited and most cited since 2024:
- You Only Look Once: Unified, Real-Time Object Detection (CVPR 2016 · 51,511 citations)
- Feature Pyramid Networks for Object Detection (CVPR 2017 · 29,987 citations)
- DETRs Beat YOLOs on Real-time Object Detection (CVPR 2024 · 4,429 citations)
- YOLOv10: Real-Time End-to-End Object Detection (NeurIPS 2024 · 1,586 citations)
vision · imagenet · transformers · 220 papers
Approaches in this cluster:
- Convolution-transformer backbones (101 papers)
Design visual backbones mixing convolution and self-attention, including masked pretraining. - Plain and high-resolution ViTs (69 papers)
Adapt vision transformers to segmentation and detection with high-resolution and plain designs. - Efficient network training and design (37 papers)
Make CNNs compact and resource-efficient via slimming, over-parameterization, and channel attention. - Vision state space models (13 papers)
Apply Mamba-style selective scan models to efficient visual representation learning.
Most cited and most cited since 2024:
- Deep Residual Learning for Image Recognition (CVPR 2016 · 228,971 citations)
- MobileNetV2: Inverted Residuals and Linear Bottlenecks (CVPR 2018 · 26,804 citations)
- Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model (ICML 2024 · 417 citations)
- TransNeXt: Robust Foveal Visual Perception for Vision Transformers (CVPR 2024 · 401 citations)
instance · panoptic · mask · 155 papers
Approaches in this cluster:
- Boundary-aware instance segmentation (83 papers)
Refine instance masks with boundary and polygon modeling and top-down bottom-up blending. - Joint detection and instance segmentation (39 papers)
Couple object detection and instance segmentation in unified or query-based models. - Unified panoptic segmentation (33 papers)
Perform panoptic segmentation with unified training, differentiable optimization, and relational context.
Most cited and most cited since 2024:
- Path Aggregation Network for Instance Segmentation (CVPR 2018 · 9,292 citations)
- Masked-Attention Mask Transformer for Universal Image Segmentation (CVPR 2022 · 3,259 citations)
- MGMap: Mask-Guided Learning for Online Vectorized HD Map Construction (CVPR 2024 · 47 citations)
- Dual DETRs for Multi-Label Temporal Action Detection (CVPR 2024 · 32 citations)
salient · saliency · sod · 108 papers
Approaches in this cluster:
- RGB salient object detection (61 papers)
Detect salient objects with attention, edge cues, and recurrent refinement. - RGB-D salient object detection (47 papers)
Use depth cues with attention and distillation for salient object detection.
Most cited and most cited since 2024:
- BASNet: Boundary-Aware Salient Object Detection (CVPR 2019 · 1,605 citations)
- Deeply Supervised Salient Object Detection With Short Connections (CVPR 2017 · 1,206 citations)
- DFormer: Rethinking RGBD Representation Learning for Semantic Segmentation (ICLR 2024 · 16 citations)
- Samba: A Unified Mamba-based Framework for General Salient Object Detection (CVPR 2025 · 15 citations)
text · event · scene · 101 papers
Approaches in this cluster:
- Arbitrary-shape scene text detection (61 papers)
Detect curved and oriented scene text with contour, region, and localization refinement. - Spiking event-camera detection (40 papers)
Detect objects from event data using spiking and hybrid neural networks.
Most cited and most cited since 2024:
- EAST: An Efficient and Accurate Scene Text Detector (CVPR 2017 · 1,852 citations)
- Synthetic Data for Text Localisation in Natural Images (CVPR 2016 · 1,557 citations)
- Scene Adaptive Sparse Transformer for Event-based Object Detection (CVPR 2024 · 37 citations)
- SRFormer: Text Detection Transformer with Incorporated Segmentation and Regression (AAAI 2024 · 23 citations)