CVPR 2025 accepted papers on 3D vision
602 of the 2,871 papers accepted at CVPR 2025 (21.0%) are on 3D vision, compared with 23.2% at CVPR 2024. They fall into 6 topics.
See CVPR 2025 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 · 188 papers
- Structured 3D Latents for Scalable and Versatile 3D Generation (247 citations)
- Continuous 3D Perception Model with Persistent State (81 citations)
- VISTA3D: A Unified Segmentation Foundation Model For 3D Medical Imaging (69 citations)
- GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control (49 citations)
- WonderWorld: Interactive 3D Scene Generation from a Single Image (36 citations)
- CAT4D: Create Anything in 4D with Multi-View Video Diffusion Models (34 citations)
- Video-3D LLM: Learning Position-Aware Video Representation for 3D Scene Understanding (34 citations)
- Enhanced Contrastive Learning with Multi-view Longitudinal Data for Chest X-ray Report Generation (25 citations)
- MET3R: Measuring Multi-View Consistency in Generated Images (25 citations)
- MIDI: Multi-Instance Diffusion for Single Image to 3D Scene Generation (24 citations)
- Diffusion Renderer: Neural Inverse and Forward Rendering with Video Diffusion Models (23 citations)
- ReconDreamer: Crafting World Models for Driving Scene Reconstruction via Online Restoration (23 citations)
and 176 more in the interactive map.
Gaussian splatting · 147 papers
- DepthSplat: Connecting Gaussian Splatting and Depth (80 citations)
- DIFIX3D+: Improving 3D Reconstructions with Single-Step Diffusion Models (48 citations)
- PUP 3D-GS: Principled Uncertainty Pruning for 3D Gaussian Splatting (47 citations)
- FlashGS: Efficient 3D Gaussian Splatting for Large-scale and High-resolution Rendering (46 citations)
- Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives (38 citations)
- DropGaussian: Structural Regularization for Sparse-view Gaussian Splatting (36 citations)
- 3DGUT: Enabling Distorted Cameras and Secondary Rays in Gaussian Splatting (33 citations)
- MoSca: Dynamic Gaussian Fusion from Casual Videos via 4D Motion Scaffolds (33 citations)
- WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments (31 citations)
- DashGaussian: Optimizing 3D Gaussian Splatting in 200 Seconds (30 citations)
- DropoutGS: Dropping Out Gaussians for Better Sparse-view Rendering (25 citations)
- NexusGS: Sparse View Synthesis with Epipolar Depth Priors in 3D Gaussian Splatting (24 citations)
and 135 more in the interactive map.
Depth and stereo · 87 papers
- VGGT: Visual Geometry Grounded Transformer (410 citations)
- MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors (81 citations)
- Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Pass (70 citations)
- MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervision (69 citations)
- MonSter: Marry Monodepth to Stereo Unleashes Power (53 citations)
- DFormerv2: Geometry Self-Attention for RGBD Semantic Segmentation (47 citations)
- DEFOM-Stereo: Depth Foundation Model Based Stereo Matching (43 citations)
- MegaSaM: Accurate, Fast and Robust Structure and Motion from Casual Dynamic Videos (41 citations)
- MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Seconds (35 citations)
- MUSt3R: Multi-view Network for Stereo 3D Reconstruction (32 citations)
- Prompting Depth Anything for 4K Resolution Accurate Metric Depth Estimation (29 citations)
- SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos (29 citations)
and 75 more in the interactive map.
Face and shape · 80 papers
- CraftsMan3D: High-fidelity Mesh Generation with 3D Native Diffusion and Interactive Geometry Refiner (34 citations)
- SF3D: Stable Fast 3D Mesh Reconstruction with UV-unwrapping and Illumination Disentanglement (29 citations)
- Dora: Sampling and Benchmarking for 3D Shape Variational Auto-Encoders (20 citations)
- PSHuman: Photorealistic Single-image 3D Human Reconstruction using Cross-Scale Multiview Diffusion and Explicit Remeshing (19 citations)
- HRAvatar: High-Quality and Relightable Gaussian Head Avatar (16 citations)
- MeGA: Hybrid Mesh-Gaussian Head Avatar for High-Fidelity Rendering and Head Editing (16 citations)
- RGBAvatar: Reduced Gaussian Blendshapes for Online Modeling of Head Avatars (16 citations)
- Instant3dit: Multiview Inpainting for Fast Editing of 3D Objects (14 citations)
- Vid2Avatar-Pro: Authentic Avatar from Videos in the Wild via Universal Prior (14 citations)
- FATE: Full-head Gaussian Avatar with Textural Editing from Monocular Video (11 citations)
- 3D Dental Model Segmentation with Geometrical Boundary Preserving (10 citations)
- PrEditor3D: Fast and Precise 3D Shape Editing (9 citations)
and 68 more in the interactive map.
Human pose estimation · 74 papers
- WiLoR: End-to-end 3D Hand Localization and Reconstruction in-the-wild (49 citations)
- HOT3D: Hand and Object Tracking in 3D from Egocentric Multi-View Videos (35 citations)
- IDOL: Instant Photorealistic 3D Human Creation from a Single Image (31 citations)
- Reloc3r: Large-Scale Training of Relative Camera Pose Regression for Generalizable, Fast, and Accurate Visual Localization (31 citations)
- FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Views (28 citations)
- Gaze-LLE: Gaze Target Estimation via Large-Scale Learned Encoders (26 citations)
- PromptHMR: Promptable Human Mesh Recovery (26 citations)
- Any6D: Model-free 6D Pose Estimation of Novel Objects (23 citations)
- You See it, You Got it: Learning 3D Creation on Pose-Free Videos at Scale (21 citations)
- Reconstructing Humans with a Biomechanically Accurate Skeleton (14 citations)
- ProbPose: A Probabilistic Approach to 2D Human Pose Estimation (13 citations)
- GCE-Pose: Global Context Enhancement for Category-level Object Pose Estimation (12 citations)
and 62 more in the interactive map.
NeRF and radiance fields · 26 papers
- Sparse Voxels Rasterization: Real-time High-fidelity Radiance Field Rendering (16 citations)
- FrugalNeRF: Fast Convergence for Extreme Few-shot Novel View Synthesis without Learned Priors (8 citations)
- Geometry Field Splatting with Gaussian Surfels (7 citations)
- InsTaG: Learning Personalized 3D Talking Head from Few-Second Video (6 citations)
- NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction (6 citations)
- GO-N3RDet: Geometry Optimized NeRF-enhanced 3D Object Detector (5 citations)
- LookCloser: Frequency-aware Radiance Field for Tiny-Detail Scene (5 citations)
- Neural Inverse Rendering from Propagating Light (5 citations)
- PBR-NeRF: Inverse Rendering with Physics-Based Neural Fields (5 citations)
- Depth-Guided Bundle Sampling for Efficient Generalizable Neural Radiance Field Reconstruction (3 citations)
- Exploiting Deblurring Networks for Radiance Fields (3 citations)
- Flow-NeRF: Joint Learning of Geometry, Poses, and Dense Flow within Unified Neural Representations (3 citations)
and 14 more in the interactive map.