1 paper
Shimin Hu, Yuanyi Wei, Fei Zha +2
Existing 3D editing methods rely on computationally intensive scene-by-scene iterative optimization and suffer from multi-view inconsistency. We propose an effective and feed-forwa…