ICLR 2023 accepted papers on 3D vision
45 of the 1,575 papers accepted at ICLR 2023 (2.9%) are on 3D vision, compared with 1.6% at ICLR 2022. They fall into 5 topics.
See ICLR 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
NeRF and radiance fields · 10 papers
- GeneFace: Generalized and High-Fidelity Audio-Driven 3D Talking Face Synthesis (33 citations)
- DM-NeRF: 3D Scene Geometry Decomposition and Manipulation from 2D Images (26 citations)
- NeRF-SOS: Any-View Self-supervised Object Segmentation on Complex Scenes (14 citations)
- S-NeRF: Neural Radiance Fields for Street Views (13 citations)
- PAC-NeRF: Physics Augmented Continuum Neural Radiance Fields for Geometry-Agnostic System Identification (7 citations)
- Is Attention All That NeRF Needs? (6 citations)
- General Neural Gauge Fields (1 citation)
- Neural Image-based Avatars: Generalizable Radiance Fields for Human Avatar Modeling (1 citation)
- Neural Radiance Field Codebooks
- Switch-NeRF: Learning Scene Decomposition with Mixture of Experts for Large-scale Neural Radiance Fields
Face and shape · 10 papers
- EVA3D: Compositional 3D Human Generation from 2D Image Collections (35 citations)
- Voxurf: Voxel-based Efficient and Accurate Neural Surface Reconstruction (31 citations)
- MeshDiffusion: Score-based Generative 3D Mesh Modeling (30 citations)
- Deep Learning on Implicit Neural Representations of Shapes (8 citations)
- Equivariant Shape-Conditioned Generation of 3D Molecules for Ligand-Based Drug Design (8 citations)
- ISS: Image as Stepping Stone for Text-Guided 3D Shape Generation (4 citations)
- VoGE: A Differentiable Volume Renderer using Gaussian Ellipsoids for Analysis-by-Synthesis (4 citations)
- Neural Implicit Shape Editing using Boundary Sensitivity (2 citations)
- Light Sampling Field and BRDF Representation for Physically-based Neural Rendering (1 citation)
- StyleMorph: Disentangled 3D-Aware Image Synthesis with a 3D Morphable StyleGAN
Human pose estimation · 9 papers
- ManiSkill2: A Unified Benchmark for Generalizable Manipulation Skills (22 citations)
- Self-Supervised Geometric Correspondence for Category-Level 6D Object Pose Estimation in the Wild (8 citations)
- Proactive Multi-Camera Collaboration for 3D Human Pose Estimation (6 citations)
- Capturing the Motion of Every Joint: 3D Human Pose and Shape Estimation with Independent Tokens (4 citations)
- Image to Sphere: Learning Equivariant Features for Efficient Pose Prediction (3 citations)
- Self-Supervised Category-Level Articulated Object Pose Estimation with Part-Level SE(3) Equivariance (3 citations)
- Rethinking Self-Supervised Visual Representation Learning in Pre-training for 3D Human Pose and Shape Estimation (1 citation)
- Decompose to Generalize: Species-Generalized Animal Pose Estimation
- LiftedCL: Lifting Contrastive Learning for Human-Centric Perception
3D generation · 9 papers
- Uni-Mol: A Universal 3D Molecular Representation Learning Framework (333 citations)
- Novel View Synthesis with Diffusion Models (62 citations)
- Autoencoders as Cross-Modal Teachers: Can Pretrained 2D Image Transformers Help 3D Representation Learning? (20 citations)
- Molecular Geometry Pretraining with SE(3)-Invariant Denoising Distance Matching (19 citations)
- Neural Groundplans: Persistent Neural Scene Representations from a Single Image (7 citations)
- One Transformer Can Understand Both 2D & 3D Molecular Data (5 citations)
- Stochastic Multi-Person 3D Motion Forecasting (3 citations)
- DualAfford: Learning Collaborative Visual Affordance for Dual-gripper Manipulation (2 citations)
- 3D generation on ImageNet
Depth and stereo · 7 papers
- Time Will Tell: New Outlooks and A Baseline for Temporal Multi-View 3D Object Detection (70 citations)
- VA-DepthNet: A Variational Approach to Single Image Depth Prediction (18 citations)
- DENSE RGB SLAM WITH NEURAL IMPLICIT MAPS (11 citations)
- Monocular Scene Reconstruction with 3D SDF Transformers (4 citations)
- LDMIC: Learning-based Distributed Multi-view Image Coding (2 citations)
- Learning to Estimate Single-View Volumetric Flow Motions without 3D Supervision
- WiNeRT: Towards Neural Ray Tracing for Wireless Channel Modelling and Differentiable Simulations