1 paper
Jinghao Hu, Yuhe Zhang, GuoHua Geng +2
Generating multi-frame, action-rich visual narratives without fine-tuning faces a threefold tension: action text faithfulness, subject identity fidelity, and cross-frame background…