Towards A Better Metric for Text-to-Video Generation

Project Page | Paper | TVGE Dataset

Jay Zhangjie Wu $^{1*}$, Guian Fang $^{1*}$, Haoning Wu $^{4*}$, Xintao Wang $^3$, Yixiao Ge $^3$, Xiaodong Cun $^3$, David Junhao Zhang $^1$, Jia-Wei Liu $^1$, Yuchao Gu $^1$, Rui Zhao $^1$, Weisi Lin $^4$, Wynne Hsu $^2$, Ying Shan $^3$, Mike Zheng Shou $^1$.<br> (* equal contribution)<br> $^1$ Show Lab, $^2$ National University of Singapore | $^3$ ARC Lab, Tencent PCG | $^4$ Nanyang Technological University

News

• [10 Apr. 2024] TVGE dataset is released. Check it out here. Code will come soon.

T2VScore-A (Text Alignment)

<img src="https://showlab.github.io/T2VScore/assets/text_alignment.gif" width="480px"/>

T2VScore-Q (Video Quality)

<img src="https://showlab.github.io/T2VScore/assets/video_quality.jpg" width="480px"/>