ECCV 2022 accepted papers on Detection and segmentation
228 of the 1,645 papers accepted at ECCV 2022 (13.9%) are on Detection and segmentation, compared with 15.3% at ECCV 2020. They fall into 5 topics.
See ECCV 2022 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Open-set and few-shot detection · 83 papers
- Detecting Twenty-Thousand Classes Using Image-Level Supervision (660 citations)
- Extract Free Dense Labels from CLIP (613 citations)
- Scaling Open-Vocabulary Image Segmentation with Image-Level Labels (473 citations)
- A Simple Baseline for Open-Vocabulary Semantic Segmentation with Pre-trained Vision-Language Model (330 citations)
- Open-Vocabulary DETR with Conditional Matching (261 citations)
- Simple Open-Vocabulary Object Detection with Vision Transformers (253 citations)
- Language-Grounded Indoor 3D Semantic Segmentation in the Wild (214 citations)
- PromptDet: Towards Open-Vocabulary Detection Using Uncurated Images (198 citations)
- Highly Accurate Dichotomous Image Segmentation (151 citations)
- ESS: Learning Event-Based Semantic Segmentation from Still Images (115 citations)
- Pixel-Wise Energy-Biased Abstention Learning for Anomaly Segmentation on Complex Urban Driving Scenes (107 citations)
- "PartImageNet: A Large, High-Quality Dataset of Parts" (105 citations)
and 71 more in the interactive map.
CNN features · 70 papers
- Exploring Plain Vision Transformer Backbones for Object Detection (1,044 citations)
- RFLA: Gaussian Receptive Field Based Label Assignment for Tiny Object Detection (378 citations)
- Multimodal Object Detection via Probabilistic Ensembling (278 citations)
- An End-to-End Transformer Model for Crowd Localization (196 citations)
- UIA-ViT: Unsupervised Inconsistency-Aware Method Based on Vision Transformer for Face Forgery Detection (163 citations)
- GRIT: Faster and Better Image Captioning Transformer Using Dual Visual Features (159 citations)
- "Translation, Scale and Rotation: Cross-Modal Alignment Meets RGB-Infrared Vehicle Detection" (156 citations)
- Cost Aggregation with 4D Convolutional Swin Transformer for Few-Shot Segmentation (152 citations)
- k-Means Mask Transformer (123 citations)
- SPSN: Superpixel Prototype Sampling Network for RGB-D Salient Object Detection (122 citations)
- ObjectBox: From Centers to Boxes for Anchor-Free Object Detection (92 citations)
- OSFormer: One-Stage Camouflaged Instance Segmentation with Transformers (91 citations)
and 58 more in the interactive map.
Weakly supervised detection and segmentation · 29 papers
- PseCo: Pseudo Labeling and Consistency Training for Semi-Supervised Object Detection (169 citations)
- Dense Teacher: Dense Pseudo-Labels for Semi-Supervised Object Detection (159 citations)
- Exploiting Unlabeled Data with Vision and Language Models for Object Detection (107 citations)
- Box-Supervised Instance Segmentation with Level Set Evolution (92 citations)
- Semi-Supervised 3D Object Detection with Proficient Teachers (87 citations)
- Max Pooling with Vision Transformers Reconciles Class and Shape in Weakly Supervised Semantic Segmentation (83 citations)
- Point-to-Box Network for Accurate Object Detection via Single Point Supervision (81 citations)
- Adversarial Erasing Framework via Triplet with Gated Pyramid Pooling Layer for Weakly Supervised Semantic Segmentation (65 citations)
- Contrastive Deep Supervision (57 citations)
- Object Discovery via Contrastive Learning for Weakly Supervised Object Detection (45 citations)
- Weakly Supervised Object Localization via Transformer with Implicit Spatial Calibration (45 citations)
- Adaptive Spatial-BCE Loss for Weakly Supervised Semantic Segmentation (44 citations)
and 17 more in the interactive map.
3D detection and driving · 26 papers
- BEVFormer: Learning Bird’s-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers (1,685 citations)
- PETR: Position Embedding Transformation for Multi-View 3D Object Detection (776 citations)
- PillarNet: Real-Time and High-Performance Pillar-Based 3D Object Detection (278 citations)
- PersFormer: 3D Lane Detection via Perspective Transformer and the OpenLane Benchmark (221 citations)
- Deformable Feature Aggregation for Dynamic Multi-modal 3D Object Detection (149 citations)
- CODA: A Real-World Road Corner Case Dataset for Object Detection in Autonomous Driving (131 citations)
- Graph R-CNN: Towards Accurate 3D Object Detection with Semantic-Decorated Local Graph (104 citations)
- Cross-Modality Knowledge Distillation Network for Monocular 3D Object Detection (99 citations)
- DID-M3D: Decoupling Instance Depth for Monocular 3D Object Detection (92 citations)
- DEVIANT: Depth EquiVarIAnt NeTwork for Monocular 3D Object Detection (90 citations)
- Monocular 3D Object Detection with Depth from Motion (68 citations)
- CramNet: Camera-Radar Fusion with Ray-Constrained Cross-Attention for Robust 3D Object Detection (64 citations)
and 14 more in the interactive map.
Tracking · 20 papers
- ByteTrack: Multi-Object Tracking by Associating Every Detection Box (3,110 citations)
- Joint Feature Learning and Relation Modeling for Tracking: A One-Stream Framework (996 citations)
- MOTR: End-to-End Multiple-Object Tracking with TRansformer (842 citations)
- Backbone Is All Your Need: A Simplified Architecture for Visual Object Tracking (346 citations)
- Towards Grand Unification of Object Tracking (177 citations)
- "FEAR: Fast, Efficient, Accurate and Robust Visual Tracker" (103 citations)
- Tracking Objects As Pixel-Wise Distributions (60 citations)
- Towards Sequence-Level Training for Visual Tracking (53 citations)
- PolarMOT: How Far Can Geometric Relations Take Us in 3D Multi-Object Tracking? (50 citations)
- Tracking Every Thing in the Wild (45 citations)
- The Caltech Fish Counting Dataset: A Benchmark for Multiple-Object Tracking and Counting (38 citations)
- Hierarchical Feature Embedding for Visual Tracking (23 citations)
and 8 more in the interactive map.
