1 paper
Yuqi Tang, Tengfei Liu, Yizheng Lai +18
Video generation increasingly relies on keyframe-based workflows, where creators specify a sequence of reference images to guide generation. Although recent models support multi-ke…