1 paper
Seungwoo Yoon, Jinmo Kim, Jaesik Park
In this paper, we propose Extend3D, a training-free pipeline for 3D scene generation from a single image, built upon an object-centric 3D generative model. To overcome the limitati…