Ziwen Chen
@chenziwee
PhD candidate @ OregonState | Computer Vision, 3D Reconstruction, Scene Understanding
Amazing work from Ziqiao! The cleanest GS-based dynamic object reconstruction solution so far
Can we scale 4D pretraining to learn general space-time representations that reconstruct an object from a few views at any time to any view at any other time? Introducing 4D-LRM: a Large Space-Time Reconstruction Model that ... 🔹 Predicts 4D Gaussian primitives directly from…
Great work! Further boosted performance of Long-LRM😆
💥 Think more real data is needed for scene reconstruction? Think again! Meet MegaSynth: scaling up feed-forward 3D scene reconstruction with synthesized scenes. In 3 days, it generates 700K scenes for training—70x larger than real data! ✨ The secret? Reconstruction is mostly…
Novel view synthesis has long been a core challenge in 3D vision. But how much 3D inductive bias is truly needed? —Surprisingly, very little! Introducing "LVSM: A Large View Synthesis Model with Minimal 3D Inductive Bias"—a fully transformer-based approach that enables scalable,…
167!
Want to find those super far away cars which are butchered by too much downsampling? Check out end-to-end adaptive downsampling! Even with 32x stride we kept enough information to segment small objects! Check us out this morning at poster 167! #CVPR2023 tinyurl.com/mr9pj62j