1 paper
Guangyu Wang, Jinzhi Zhang, Fan Wang +2
We propose XScale-NVS for high-fidelity cross-scale novel view synthesis of real-world large-scale scenes. Existing representations based on explicit surface suffer from discretiza…