NeurIPS 2023 accepted papers on 3D vision
120 of the 3,584 papers accepted at NeurIPS 2023 (3.3%) are on 3D vision, compared with 3.2% at NeurIPS 2022. They fall into 5 topics.
See NeurIPS 2023 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
3D generation · 34 papers
- Objaverse-XL: A Universe of 10M+ 3D Objects (97 citations)
- Occ3D: A Large-Scale 3D Occupancy Prediction Benchmark for Autonomous Driving (55 citations)
- SceneScape: Text-Driven Consistent Scene Generation (24 citations)
- Scalable 3D Captioning with Pretrained Models (21 citations)
- OpenLane-V2: A Topology Reasoning Benchmark for Unified 3D HD Mapping (18 citations)
- CommonScenes: Generating Commonsense 3D Indoor Scenes with Scene Graph Diffusion (13 citations)
- Stanford-ORB: A Real-World 3D Object Inverse Rendering Benchmark (10 citations)
- Autodecoding Latent 3D Diffusion Models (9 citations)
- NAP: Neural 3D Articulated Object Prior (8 citations)
- DDF-HO: Hand-Held Object Reconstruction via Conditional Directed Distance Field (7 citations)
- Towards A Richer 2D Understanding of Hands at Scale (7 citations)
- AV-NeRF: Learning Neural Fields for Real-World Audio-Visual Scene Synthesis (6 citations)
and 22 more in the interactive map.
NeRF and radiance fields · 26 papers
- Segment Anything in 3D with NeRFs (34 citations)
- ViCA-NeRF: View-Consistency-Aware 3D Editing of Neural Radiance Fields (17 citations)
- Binary Radiance Fields (6 citations)
- Inferring Hybrid Neural Fluid Fields from Videos (6 citations)
- Transient Neural Radiance Fields for Lidar View Synthesis and 3D Reconstruction (6 citations)
- DäRF: Boosting Radiance Fields from Sparse Input Views with Monocular Depth Adaptation (5 citations)
- FlowCam: Training Generalizable 3D Radiance Fields without Camera Poses via Pixel-Aligned Scene Flow (5 citations)
- NVFi: Neural Velocity Fields for 3D Physics Learning from Dynamic Videos (5 citations)
- PanoGRF: Generalizable Spherical Radiance Fields for Wide-baseline Panoramas (4 citations)
- DynPoint: Dynamic Neural Point For View Synthesis (3 citations)
- Mip-Grid: Anti-aliased Grid Representations for Neural Radiance Fields (3 citations)
- ConRad: Image Constrained Radiance Fields for 3D Generation from a Single Image (2 citations)
and 14 more in the interactive map.
Face and shape · 25 papers
- One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization (107 citations)
- DreamHuman: Animatable 3D Avatars from Text (22 citations)
- DreamWaltz: Make a Scene with Complex 3D Animatable Avatars (16 citations)
- DiffComplete: Diffusion-based Generative 3D Shape Completion (11 citations)
- ARTIC3D: Learning Robust Articulated 3D Shapes from Noisy Web Image Collections (10 citations)
- DELIFFAS: Deformable Light Fields for Fast Avatar Synthesis (8 citations)
- Global-correlated 3D-decoupling Transformer for Clothed Avatar Reconstruction (8 citations)
- NAVI: Category-Agnostic Image Collections with High-Quality 3D Shape and Pose Annotations (8 citations)
- Jigsaw: Learning to Assemble Multiple Fractured Objects (6 citations)
- Coupled Reconstruction of Cortical Surfaces by Diffeomorphic Mesh Deformation (5 citations)
- ISP: Multi-Layered Garment Draping with Implicit Sewing Patterns (5 citations)
- RenderMe-360: A Large Digital Asset Library and Benchmarks Towards High-fidelity Head Avatars (5 citations)
and 13 more in the interactive map.
Depth and stereo · 18 papers
- Dynamo-Depth: Fixing Unsupervised Depth Estimation for Dynamical Scenes (11 citations)
- Fine-Grained Cross-View Geo-Localization Using a Correlation-Aware Homography Estimator (11 citations)
- EV-Eye: Rethinking High-frequency Eye Tracking through the Lenses of Event Cameras (8 citations)
- Deep Non-line-of-sight Imaging from Under-scanning Measurements (7 citations)
- Learning Dense Flow Field for Highly-accurate Cross-view Camera Localization (6 citations)
- GenS: Generalizable Neural Surface Reconstruction from Multi-View Images (5 citations)
- OpenIllumination: A Multi-Illumination Dataset for Inverse Rendering Evaluation on Real Objects (5 citations)
- Hyper-Skin: A Hyperspectral Dataset for Reconstructing Facial Skin-Spectra from RGB Images (4 citations)
- Lookup Table meets Local Laplacian Filter: Pyramid Reconstruction Network for Tone Mapping (3 citations)
- Learning Neural Implicit through Volume Rendering with Attentive Depth Fusion Priors (2 citations)
- MVDoppler: Unleashing the Power of Multi-View Doppler for MicroMotion-based Gait Classification (2 citations)
- NeRF-IBVS: Visual Servo Based on NeRF for Visual Localization and Navigation (1 citation)
and 6 more in the interactive map.
Human pose estimation · 17 papers
- SMPLer-X: Scaling Up Expressive Human Pose and Shape Estimation (29 citations)
- Motion-X: A Large-scale 3D Expressive Whole-body Human Motion Dataset (23 citations)
- Generative Category-level Object Pose Estimation via Diffusion Models (13 citations)
- A Single 2D Pose with Context is Worth Hundreds for 3D Human Pose Estimation (8 citations)
- Towards Robust and Expressive Whole-body Human Pose and Shape Estimation (7 citations)
- HAP: Structure-Aware Masked Image Modeling for Human-Centric Perception (5 citations)
- A Dataset of Relighted 3D Interacting Hands (4 citations)
- H-InDex: Visual Reinforcement Learning with Hand-Informed Representations for Dexterous Manipulation (4 citations)
- DeepSimHO: Stable Pose Estimation for Hand-Object Interaction via Physics Simulation (3 citations)
- Triangulation Residual Loss for Data-efficient 3D Pose Estimation (3 citations)
- Humans in Kitchens: A Dataset for Multi-Person Human Motion Forecasting with Scene Context (2 citations)
- Equivariant Single View Pose Prediction Via Induced and Restriction Representations (1 citation)
and 5 more in the interactive map.