Skip to content

Retrieval-Augmented Video Generation for Telling a Story

Notifications You must be signed in to change notification settings

AILab-CVC/Animate-A-Story

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Animate-A-Story: Storytelling with Retrieval-Augmented Video Generation

     

Yingqing He*, Menghan Xia*, Haoxin Chen*, Xiaodong Cun, Yuan Gong, Jinbo Xing,
Yong Zhang#, Xintao Wang, Chao Weng, Ying Shan and Qifeng Chen#

(* first author, # corresponding author)

🥳 Demo

Please check more demo videos at the project page.

🔆 Abstract

TL; DR: 🤗🤗🤗 Animate-A-Story is a video storytelling approach which can synthesize high-quality, structure-controlled, and character-controlled videos.

Generating videos for visual storytelling can be a tedious and complex process that typically requires either live-action filming or graphics animation rendering. To bypass these challenges, our key idea is to utilize the abundance of existing video clips and synthesize a coherent storytelling video by customizing their appearances. We achieve this by developing a framework comprised of two functional modules: (i) Motion Structure Retrieval, which provides video candidates with desired scene or motion context described by query texts, and (ii) Structure-Guided Text-to-Video Synthesis, which generates plot-aligned videos under the guidance of motion structure and text prompts. For the first module, we leverage an off-the-shelf video retrieval system and extract video depths as motion structure. For the second module, we propose a controllable video generation model that offers flexible controls over structure and characters. The videos are synthesized by following the structural guidance and appearance instruction. To ensure visual consistency across clips, we propose an effective concept personalization approach, which allows the specification of the desired character identities through text prompts. Our experiments showcase the significant advantages of our proposed methods over various existing baselines. Moreover, user studies on our synthesized storytelling videos demonstrate the effectiveness of our framework and indicate the promising potential for practical applications.

😉 Citation

@article{he2023animate,
  title={Animate-a-story: Storytelling with retrieval-augmented video generation},
  author={He, Yingqing and Xia, Menghan and Chen, Haoxin and Cun, Xiaodong and Gong, Yuan and Xing, Jinbo and Zhang, Yong and Wang, Xintao and Weng, Chao and Shan, Ying and others},
  journal={arXiv preprint arXiv:2307.06940},
  year={2023}
}

📭 Contact

If your have any comments or questions, feel free to contact Yingqing He, Menghan Xia or Haoxin Chen.

About

Retrieval-Augmented Video Generation for Telling a Story

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published