Pixie: Physics from Pixels

Photorealistic 3D reconstructions (NeRF, Gaussian Splatting) capture geometry & appearance but lack physics. This limits 3D reconstruction to static scenes. Recently, there has been a surge of interest in integrating physics into 3D modeling. But existing test‑time optimisation methods are slow and scene‑specific. Pixie trains a neural network thatmaps pretrained visual features (i.e., CLIP) to dense material fields of physical properties in a single forward pass, enabling real‑time physics simulations.