AlignVid: Training-Free Attention Scaling for Semantic Fidelity in Text-Guided Image-to-Video Generation
Yexin Liu
Wenjie Shu
Zile Huang
Haoze Zheng
Yueze Wang
Jingjin Zhu
Manyuan Zhang
Ser-Nam Lim
Harry Yang
Abstract
Training-free attention scaling method to improve semantic fidelity in text-guided image-to-video generation.
Type
Publication
ICML 2026
Accepted at ICML 2026.
