video-generation-dataset
[CVPR 2025 Highlightš„] Identity-Preserving Text-to-Video Generation by Frequency Decomposition