CVPR 2017 accepted papers on 3D vision
139 of the 783 papers accepted at CVPR 2017 (17.8%) are on 3D vision, compared with 19.6% at CVPR 2016. They fall into 4 topics.
See CVPR 2017 in the interactive map
Submitting to CVPR? 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
Depth and stereo · 63 papers
- Unsupervised Monocular Depth Estimation With Left-Right Consistency (3,325 citations)
- Unsupervised Learning of Depth and Ego-Motion From Video (2,885 citations)
- Deep Multi-Scale Convolutional Neural Network for Dynamic Scene Deblurring (2,358 citations)
- Semantic Scene Completion From a Single Depth Image (1,301 citations)
- Switching Convolutional Neural Network for Crowd Counting (1,018 citations)
- A Multi-View Stereo Benchmark With High-Resolution Images and Multi-Camera Videos (913 citations)
- Geometric Loss Functions for Camera Pose Regression With Deep Learning (815 citations)
- CNN-SLAM: Real-Time Dense Monocular SLAM With Learned Depth Prediction (775 citations)
- DeMoN: Depth and Motion Network for Learning Monocular Stereo (770 citations)
- Semi-Supervised Deep Learning for Monocular Depth Map Prediction (692 citations)
- Multi-Scale Continuous CRFs as Sequential Deep Networks for Monocular Depth Estimation (490 citations)
- Event-Based Visual Inertial Odometry (257 citations)
and 51 more in the interactive map.
Human pose estimation · 32 papers
- Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields (7,514 citations)
- Learning From Synthetic Humans (974 citations)
- Coarse-To-Fine Volumetric Prediction for Single-Image 3D Human Pose (947 citations)
- Multi-Context Attention for Human Pose Estimation (739 citations)
- 3D Human Pose Estimation = 2D Pose Estimation + Matching (604 citations)
- Multiple People Tracking by Lifted Multicut and Person Re-Identification (581 citations)
- Unite the People: Closing the Loop Between 3D and 2D Human Representations (572 citations)
- Lifting From the Deep: Convolutional 3D Pose Estimation From a Single Image (547 citations)
- Deep MANTA: A Coarse-To-Fine Many-Task Network for Joint 2D and 3D Vehicle Analysis From Monocular Image (484 citations)
- 3D Human Pose Estimation From a Single Image via Distance Matrix Regression (424 citations)
- LCR-Net: Localization-Classification-Regression for Human Pose (317 citations)
- 3D Convolutional Neural Networks for Efficient and Robust Hand Pose Estimation From Single Depth Images (311 citations)
and 20 more in the interactive map.
Face and shape · 30 papers
- 3DMatch: Learning Local Geometric Descriptors From RGB-D Reconstructions (1,155 citations)
- Shape Completion Using 3D-Encoder-Predictor CNNs and Shape Synthesis (678 citations)
- Multi-View Supervision for Single-View Reconstruction via Differentiable Ray Consistency (571 citations)
- Regressing Robust and Discriminative 3D Morphable Models With a Very Deep Neural Network (534 citations)
- SyncSpecCNN: Synchronized Spectral CNN for 3D Shape Segmentation (475 citations)
- 3D Shape Segmentation With Projective Convolutional Networks (395 citations)
- 3D Menagerie: Modeling the 3D Shape and Pose of Animals (388 citations)
- Dynamic FAUST: Registering Human Bodies in Motion (378 citations)
- Learning Detailed Face Reconstruction From a Single Image (376 citations)
- Learning Shape Abstractions by Assembling Volumetric Primitives (346 citations)
- End-To-End 3D Face Reconstruction With Deep Neural Networks (273 citations)
- Synthesizing 3D Shapes via Modeling Multi-View Depth Maps and Silhouettes With Deep Generative Networks (202 citations)
and 18 more in the interactive map.
3D generation · 14 papers
- ScanNet: Richly-Annotated 3D Reconstructions of Indoor Scenes (4,275 citations)
- Hand Keypoint Detection in Single Images Using Multiview Bootstrapping (1,182 citations)
- Latent Multi-View Subspace Clustering (573 citations)
- Transformation-Grounded Image Generation Network for Novel 3D View Synthesis (293 citations)
- IM2CAD (155 citations)
- Neural Scene De-Rendering (110 citations)
- Deep Supervision With Shape Concepts for Occlusion-Aware 3D Object Parsing (91 citations)
- Binge Watching: Scaling Affordance Learning From Sitcoms (81 citations)
- Shading Annotations in the Wild (76 citations)
- Semantically Coherent Co-Segmentation and Reconstruction of Dynamic Scenes (47 citations)
- Exploiting 2D Floorplan for Building-Scale Panorama RGBD Alignment (42 citations)
- DUST: Dual Union of Spatio-Temporal Subspaces for Monocular Multiple Object 3D Reconstruction (32 citations)
and 2 more in the interactive map.
