ICCV 2021 accepted papers on Point clouds
102 of the 1,612 papers accepted at ICCV 2021 (6.3%) are on Point clouds, compared with 4.0% at ICCV 2019. They fall into 4 topics.
See ICCV 2021 in the interactive map
Submitting to ICCV? Checklist →
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
Point cloud classification · 37 papers
- Point Transformer (3,038 citations)
- Walk in the Cloud: Learning Curves for Point Clouds Shape Analysis (341 citations)
- Unsupervised Point Cloud Pre-Training via Occlusion Completion (283 citations)
- Adaptive Graph Convolution for Point Cloud Analysis (211 citations)
- Spatio-Temporal Self-Supervised Representation Learning for 3D Point Clouds (207 citations)
- 3DVG-Transformer: Relation Modeling for Visual Grounding on Point Clouds (161 citations)
- Pyramid Point Cloud Transformer for Large-Scale Place Recognition (153 citations)
- InstanceRefer: Cooperative Holistic Understanding for Visual Grounding on Point Clouds Through Instance Multi-Level Contextual Referring (129 citations)
- Box-Aware Feature Enhancement for Single Object Tracking on Point Clouds (125 citations)
- SAT: 2D Semantics Assisted Training for 3D Visual Grounding (112 citations)
- Unsupervised Depth Completion With Calibrated Backprojection Layers (102 citations)
- GANcraft: Unsupervised 3D Neural Rendering of Minecraft Worlds (99 citations)
and 25 more in the interactive map.
Segmentation and detection · 26 papers
- Voxel Transformer for 3D Object Detection (542 citations)
- RPVNet: A Deep and Efficient Range-Point-Voxel Fusion Network for LiDAR Point Cloud Segmentation (370 citations)
- Group-Free 3D Object Detection via Transformers (318 citations)
- Self-Supervised Pretraining of 3D Features on Any Point-Cloud (310 citations)
- Improving 3D Object Detection With Channel-Wise Transformer (307 citations)
- Fog Simulation on Real LiDAR Point Clouds for 3D Object Detection in Adverse Weather (233 citations)
- Perception-Aware Multi-Sensor Fusion for 3D LiDAR Semantic Segmentation (231 citations)
- Pyramid R-CNN: Towards Better Performance and Adaptability for 3D Object Detection (193 citations)
- Hierarchical Aggregation for 3D Instance Segmentation (188 citations)
- Perturbed Self-Distillation: Weakly Supervised Large-Scale Point Cloud Semantic Segmentation (172 citations)
- Guided Point Contrastive Learning for Semi-Supervised Point Cloud Semantic Segmentation (138 citations)
- Instance Segmentation in 3D Scenes Using Semantic Superpoint Tree Networks (130 citations)
and 14 more in the interactive map.
Shape and surface completion · 24 papers
- PoinTr: Diverse Point Cloud Completion With Geometry-Aware Transformers (646 citations)
- 3D Shape Generation and Completion Through Point-Voxel Diffusion (482 citations)
- SnowflakeNet: Point Cloud Completion by Snowflake Point Deconvolution With Skip-Transformer (387 citations)
- DeepCAD: A Deep Generative Network for Computer-Aided Design Models (262 citations)
- Score-Based Point Cloud Denoising (227 citations)
- Voxel-Based Network for Shape Completion by Leveraging Edge Generation (86 citations)
- SA-ConvONet: Sign-Agnostic Optimization of Convolutional Occupancy Networks (78 citations)
- AdaFit: Rethinking Learning-Based Normal Estimation on Point Clouds (62 citations)
- Learning With Noisy Labels for Robust Point Cloud Segmentation (53 citations)
- PU-EVA: An Edge-Vector Based Approximation Solution for Flexible-Scale Point Cloud Upsampling (53 citations)
- CSG-Stump: A Learning Friendly CSG-Like Representation for Interpretable Shape Parsing (48 citations)
- Point-Based Modeling of Human Clothing (47 citations)
and 12 more in the interactive map.
Registration and matching · 15 papers
- Rethinking Counting and Localization in Crowds: A Purely Point-Based Framework (437 citations)
- OMNet: Learning Overlapping Mask for Partial-to-Partial Point Cloud Registration (230 citations)
- HRegNet: A Hierarchical Network for Large-Scale Outdoor LiDAR Point Cloud Registration (147 citations)
- Deep Hough Voting for Robust Global Registration (122 citations)
- PCAM: Product of Cross-Attention Matrices for Rigid Registration of Point Clouds (70 citations)
- LSG-CPD: Coherent Point Drift With Local Surface Geometry for Point Cloud Registration (49 citations)
- Sampling Network Guided Cross-Entropy Method for Unsupervised Point Cloud Registration (46 citations)
- SOMA: Solving Optical Marker-Based MoCap Automatically (46 citations)
- Feature Interactive Representation for Point Cloud Registration (37 citations)
- A Robust Loss for Point Cloud Registration (34 citations)
- DeepPRO: Deep Partial Point Cloud Registration of Objects (26 citations)
- Distinctiveness Oriented Positional Equilibrium for Point Cloud Registration (22 citations)
and 3 more in the interactive map.
