Di Huang

ND-SDF: Learning Normal Deflection Fields for High-Fidelity Indoor Reconstruction

Depth Any Video with Scalable Synthetic Data

Where Am I and What Will I See: An Auto-Regressive Model for Spatial Localization and View Prediction

MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformers

NeuRodin: A Two-stage Framework for High-Fidelity Neural Surface Reconstruction

UniPAD: A Universal Pre-training Paradigm for Autonomous Driving