creative-graphic-design/PKU-PosterLayout

Dataset Card for PKU-PosterLayout

 

Dataset Description

• —Homepage: http://59.108.48.34/tiki/PosterLayout/
• —Repository: https://github.com/creative-graphic-design/huggingface-datasets/tree/main/datasets/PKUPosterLayout
• —Hugging Face Dataset: https://huggingface.co/datasets/creative-graphic-design/PKU-PosterLayout
• —Paper (arXiv): https://arxiv.org/abs/2303.15937
• —Paper (CVPR 2023): https://openaccess.thecvf.com/content/CVPR2023/html/HsuPosterLayoutANewBenchmarkandApproachforContent-AwareVisual-TextualPresentationCVPR2023_paper.html

Dataset Summary

PKU-PosterLayout is a content-aware visual-textual poster layout benchmark released with PosterLayout: A New Benchmark and Approach for Content-aware Visual-Textual Presentation Layout. The paper defines the task as arranging predefined text, logo, and underlay elements on a non-empty poster canvas while considering both inter-element and inter-layer relationships. The original benchmark contains 9,974 poster-layout pairs and 905 non-empty canvas images.

This Hugging Face release exposes the repository loader output as parquet files. It includes the original loader-style default config and a ralf-style config for downstream layout-generation workflows.

Supported Tasks and Leaderboards

The dataset supports poster layout generation and layout-conditioned image editing. No public leaderboard is bundled with this Hugging Face packaging.

Languages

Poster text is primarily Chinese (zh).

Dataset Structure

Data Fields

The default config contains original_poster, inpainted_poster, basnet_saliency_map, pfpn_saliency_map, canvas, and annotations.

The ralf-style config contains image_id, original_poster, inpainted_poster, canvas, saliency_map, saliency_map_sub, and annotations.

Data Splits

Dataset Creation

The original dataset extends PosterErase with layout annotations for content-aware poster layout generation. The annotations describe visual-textual elements and their positions on poster canvases.

Considerations for Using the Data

The dataset is focused on Chinese poster layouts and may not represent other languages, writing systems, or design cultures. Images and poster text remain subject to the original dataset terms.

Additional Information

Licensing Information

Images in PKU-PosterLayout are distributed under the CC BY-SA 4.0 license according to the local loader metadata.

Citation Information

@inproceedings{hsu2023posterlayout,
  title={PosterLayout: A New Benchmark and Approach for Content-aware Visual-Textual Presentation Layout},
  author={Hsu, Hsiao Yuan and He, Xiangteng and Peng, Yuxin and Kong, Hao and Zhang, Qing},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={6018--6026},
  year={2023}
}

Contributions

Thanks to PKU-ICST-MIPL for creating the original dataset.