Reproduce “VSD2M: A Large-scale Vision-language Sticker Dataset for Multi-frame Animated Sticker Generation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhiqiang Yuan, Jiapei Zhang, Ying Deng, Yeshuang Zhu, Jie Zhou, Jinchao Zhang
VSD2M introduces a large animated-sticker dataset with action-focused descriptions and adapts a video model by replacing its temporal layers with STI, which combines cross-frame semantic interaction with detail-preserving processing. In cartoon-domain tests using eight frames at 256 × 256, the modified model outperforms VideoLDM on VQA and FVD, but I2VGen-XL has higher CLIP similarity, and the STI components were not separately tested.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.