ECCV 2026 accepted papers on 3D vision
620 of the 2,834 papers accepted at ECCV 2026 (21.9%) are on 3D vision, compared with 23.0% at ECCV 2024. They fall into 6 topics.
See ECCV 2026 in the interactive map
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
3D generation · 188 papers
- Ultra3D: Efficient and High-Fidelity 3D Generation with Part Attention (51 citations)
- Skyfall-GS: Synthesizing Immersive 3D Urban Scenes from Satellite Imagery (28 citations)
- TabletopGen: Tabletop Scene Generation and Interactive Simulation for Robotic Manipulation (16 citations)
- Edit in 2D, Verify in 3D: Reinforcement Learning for Multi-view Consistent Scene Editing (14 citations)
- Repurposing Geometric Foundation Models for Multi-view Diffusion (12 citations)
- AnyView: Synthesizing Any Novel View in Dynamic Scenes (11 citations)
- One Video, One World: Turning Monocular Video into Physical 4D Scenes (11 citations)
- Diffusion Models are Open-World Affordance Learners: Leveraging Generative Priors for 3D Affordance Learning (9 citations)
- GENA3D: Generative Amodal 3D Modeling by Bridging 2D Priors and 3D Coherence (7 citations)
- Egocentric World Model for Photorealistic Hand Object Interaction Synthesis (6 citations)
- EgoCogNav: Cognition-aware Human Egocentric Navigation (6 citations)
- ECHO: Ego-centric Modeling of Human-Object Interactions (5 citations)
and 176 more in the interactive map.
Gaussian splatting · 170 papers
- VolSplat: Rethinking Feed-Forward 3D Gaussian Splatting with Voxel-Aligned Prediction (49 citations)
- MVGS: Multi-view Regulated Gaussian Splatting for Novel View Synthesis (28 citations)
- GlobalSplat: Efficient Feed-Forward 3D Gaussian Splatting via Global Scene Tokens (12 citations)
- LiteGS: a high-performance framework to train 3dgs in subminutes via system and algorithm codesign (11 citations)
- Habitat-GS: A High-Fidelity Navigation Simulator with Dynamic Gaussian Splatting (8 citations)
- DINO-SLAM: DINO-Informed RGB-D SLAM for Neural Implicit and Explicit Representations (7 citations)
- From Blobs to Spokes: High-Fidelity Surface Reconstruction via Oriented Gaussians (6 citations)
- SplatPainter: Interactive Authoring of 3D Gaussians from 2D Edits via Test-Time Training (6 citations)
- FLEG: Feed-Forward Language Embedded Gaussian Splatting from Any Views via Compact Semantic Representation (5 citations)
- HandSCS: Structural Coordinate Space for Animatable Hand Gaussian Splatting (5 citations)
- Predictive Photometric Uncertainty in Gaussian Splatting for Novel View Synthesis (5 citations)
- Active View Selection with Perturbed Gaussian Ensemble for Tomographic Reconstruction (4 citations)
and 158 more in the interactive map.
Depth and stereo · 121 papers
- RoMa v2: Harder Better Faster Denser Feature Matching (55 citations)
- Geometric Context Transformer for Streaming 3D Reconstruction (53 citations)
- Masked Depth Modeling for Spatial Perception (40 citations)
- Track4World: Feedforward World-centric Dense 3D Tracking of All Pixels (15 citations)
- LoMa: Local Feature Matching Revisited (13 citations)
- On Geometric Understanding and Learned Priors in Feed-forward 3D Reconstruction Models (13 citations)
- RayRoPE: Projective Ray Positional Encoding for Multi-view Attention (11 citations)
- Flow4R: Unifying 4D Reconstruction and Tracking with Scene Flow (8 citations)
- Fisheye3R: Adapting Unified 3D Feed-Forward Foundation Models to Fisheye Lenses (7 citations)
- MetricAnything: Scaling Metric Depth Pretraining with Noisy Heterogeneous Sources (7 citations)
- Any to Full: Prompting Depth Anything for Depth Completion in One Stage (6 citations)
- Wid3R: Wide Field-of-View 3D Reconstruction via Camera Model Conditioning (6 citations)
and 109 more in the interactive map.
Face and shape · 78 papers
- Human Mesh Modeling for Anny Body (17 citations)
- MonoArt: Progressive Structural Reasoning for Monocular Articulated 3D Reconstruction (8 citations)
- ReconPhys: Reconstruct Appearance and Physical Attributes from Single Video (6 citations)
- Calibrated Harmonic Overlaid Implicit Neural Representations for Multi-Dimensional Data (3 citations)
- MessyKitchens: Contact-rich object-level 3D scene reconstruction (3 citations)
- Mind-to-Face: Neural-Driven Photorealistic Avatar Synthesis via EEG Decoding (3 citations)
- TORA: Topological Representation Alignment for 3D Shape Assembly (3 citations)
- AHOY! Animatable Humans under Occlusion from YouTube Videos with Gaussian Splatting and Video Diffusion Priors (2 citations)
- MSVS-VAE: Multi-Scale Anchored VecSet for High-Fidelity 3D Reconstruction (2 citations)
- TriFlow: Generating Artist-Like 3D Mesh Topology via Nearest-Vertex Vector Fields (2 citations)
- Beyond Prompts: Unconditional 3D Inversion for Out-of-Distribution Shapes (1 citation)
- Hyper-Network Neural Functional Maps for Unsupervised Robust 3D Shape Matching (1 citation)
and 66 more in the interactive map.
Human pose estimation · 58 papers
- XYZ-IBD: Benchmarking Robust 6D Object Pose Estimation under Real-World Industrial Complexity (17 citations)
- Fast Sam 3D Body: Accelerating SAM 3D Body for Real-Time Full-Body Human Mesh Recovery (8 citations)
- Reconstruction by Generation: 3D Multi-Object Scene Reconstruction from Sparse Observations (8 citations)
- Cambrian-P: Pose-Grounded Video Understanding (7 citations)
- DETRPose: Real-Time End-to-End Multi-Person Pose Estimation via Modified Transformer Decoder and Novel Denoising Keypoints (7 citations)
- WiFi-JEPA: Self-supervised Learning for WiFi-CSI 3D Human Pose Estimation (4 citations)
- OrthoTrack: Continuous 6-DoF UAV Trajectory Estimation Anchored in Public Orthophotos (3 citations)
- CanoVerse: 3D Object Scalable Canonicalization and Dataset for Generation and Pose (2 citations)
- GRAFT: Geometric Refinement and Fitting Transformer for Human Scene Reconstruction (2 citations)
- HumanTracker: Towards Comprehensive and Human-Aligned Motion Tracking Benchmark (2 citations)
- Point2Pose: Occlusion-Recovering 6D Pose Tracking and 3D Reconstruction for Multiple Unknown Objects Via 2D Point Trackers (2 citations)
- SKEL-CF: Coarse-to-Fine Biomechanical Skeleton and Surface Mesh Recovery (2 citations)
and 46 more in the interactive map.
NeRF and radiance fields · 5 papers
- Do Flat Minima Improve Sparse Novel View Synthesis?
- High-speed Imaging through Turbulence with Event-based Light Fields
- Learning Spectral and Polarimetric Clues for One-to-Multimodal Novel View Synthesis
- SkyLume: A Large-Scale Multi-Illumination Aerial Benchmark for Urban Scene Reconstruction and Beyond
- OmniRen: Neural Rendering wih Heterogeneous Scene Primitives
