no code implementations • 20 Nov 2023 • Tuen-Yue Tsui, Qin Zou
Contrary to previous methods that recover the geometry, material, and illumination in multiple stages and extract the properties from various multi-layer perceptrons across different neural fields, we question such complexities and introduce our method - a single-stage framework that uniformly recovers all properties.