1 paper · 1 filter
Yuancheng Xu, Wenqi Xian, Li Ma +10
We introduce a framework that enables both multi-view character consistency and 3D camera control in video diffusion models through a novel customization data pipeline. We train th…