All posters are presented in Exhibit Hall A on June 3, 2026, 5:05–6:00 PM. Please find your assigned board number below.
A Best Demo will be featured at the end of the session.
| Board | Side | Title |
|---|---|---|
| 249 | A | MoRE: 3D Visual Geometry Reconstruction Meets Mixture-of-Experts |
| 249 | B | Sky2Ground: A Benchmark for Site Modeling under Varying Altitude |
| 250 | A | LoGeR: Long-Context Geometric Reconstruction with Hybrid Memory |
| 250 | B | Generalizable Sparse-View 3D Reconstruction from Unconstrained Images |
| 251 | A | Not All Points Are Equal: Uncertainty-Aware 4D LiDAR Scene Synthesis |
| 251 | B | WildDet3D: Scaling Promptable 3D Detection in the Wild |
| 252 | A | Video2Assets: Extracting 3D Object Assets from Unconstrained Video |
| 252 | B | PAGE-4D: VGGT-4D Perception Via Disentangled Pose and Geometry Estimation |
| 253 | A | PointWorld: Scaling 3D World Models for In-The-Wild Robotic Manipulation ⭐ Best Paper |
| 253 | B | LA-Pose: Latent Action Pretraining Meets Pose Estimation |
| 254 | A | FreeOrbit4D: Training-Free Arbitrary Camera Redirection for Monocular Videos via Geometry-Complete 4D Reconstruction |
| 254 | B | FF3R: Feedforward Feature 3D Reconstruction from Unconstrained views |
| 255 | A | Long-Tail Internet Photo Reconstruction |
| 255 | B | Feedforward 3D Editing via Text-Steerable Image-to-3D |
| 256 | A | Balancing Local Geometry and Global Semantics in 3D Point Clouds with TAP-JEPA |
| 256 | B | Emergent Extreme-View Geometry in 3D Foundation Models |
| 257 | A | Trace Anything: Representing Any Video in 4D via Trajectory Fields |
| 257 | B | Dream2Flow: Bridging Video Generation and Open-World Manipulation with 3D Object Flow |
| 258 | A | Safe2Drive: Evaluating Safe Driving Behaviors of E2E Autonomous Driving Models |
| 258 | B | Sat3R: Satellite DSM Reconstruction via RPC-Aware Depth Fine-tuning |
| 259 | A | GS4: Generalizable Sparse Splatting Semantic SLAM |
| 259 | B | Self-Questioning Vision-Language Models: Reinforcement Learning for Compositional Visual Reasoning |
| 260 | A | 3D Point World Models: Point Completion Enables More Accurate and Generalizable Dynamics Learning |
| 260 | B | AuditSplat: Self-Auditing Feed-Forward 3D Gaussian Splatting for Web-Scale Auto-Annotation |
| 261 | A | Outdoor-Targeted Perceptual Specialisation in 3DGS-MCMC via Late-Stage Anisotropic Primitive Splitting at Iso-Compute |
| 261 | B | AttDis & Friends: Three Lightweight Uncertainty Channels in Frozen Feed-Forward 3D Gaussian Splatting |
| 262 | A | Changes in Real Time: Online Scene Change Detection with Multi-View Fusion |
| 262 | B | Learning Surgical Robotic Manipulation with 3D Spatial Priors ⭐ Best Transitional |
| 263 | A | The Epipolar Bottleneck: Minimal Sufficient Statistics for Sample-Efficient End-to-End 3D Reconstruction |
| 263 | B | SpectraSync: Differentiable Spectral Synchronization for End-to-End Multi-View 3D Reconstruction |
| 264 | A | The Spectral Gap as a Geometric Prior: Laplacian-Biased Attention for End-to-End Multi-View 3D Reconstruction |
| 264 | B | Geometric Sufficiency: An Information-Theoretic Framework for Task-Aware 3D Representations |
| 265 | A | Streaming3D: Sequential 3D Generation via Evidential Memory |
| 265 | B | Benchmarking stereo reconstruction for 3D printable Martian terrain models |
| 266 | A | CellSplat: End-to-End 3D Gaussian Splatting for Cellular Decomposition of Spatial Transcriptomics |