1 paper
Yitong Xia, Zeren Jiang, Chuanxia Zheng
We introduce Gen4P, a unified framework for recovering 4D geometry and motion from monocular video. We show that a pretrained video generator can be repurposed into a feed-forward…