Scal3R: Learning Efficient Multi-Relative Pose Query for Scalable Online 3D Reconstruction Paper • 2609.04201 • Published 3 days ago • 45
Marionette: Predicting World States, Rendering Geometry, Painting Appearance Paper • 2608.14530 • Published 23 days ago • 34
HelloWorld: Enabling Socially Interactive Characters in Video World Models Paper • 2608.05070 • Published Aug 5 • 40
LongE2V: Long-Horizon Event-based Video Reconstruction, Prediction, and Frame Interpolation with Video Diffusion Models Paper • 2607.08770 • Published Jul 9 • 37
BRDFusion: Physics Meets Generation for Urban Scene Inverse Rendering Paper • 2606.17049 • Published Jun 15 • 29
Reroute, Don't Remove: Recoverable Visual Token Routing for Vision-Language Models Paper • 2606.12412 • Published Jun 10 • 21