Reproduce “Calliar: An Online Handwritten Dataset for Arabic Calligraphy” 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

Zaid Alyafeai, Maged S. Al-shaibani, Mustafa Ghaleb, Yousif Ahmed Al-Wajih

Researchers at King Fahd University of Petroleum and Minerals developed Calliar, the first large-scale online handwritten dataset for Arabic calligraphy, encompassing 2,500 sentences and over 45,000 strokes across various calligraphic styles. This resource provides stroke, character, word, and sentence-level annotations, facilitating research in areas like online handwriting recognition and generative artistic script models.

Key insights

  • Arabic calligraphy presents significant variability, with characters drawn upside down or with different stroke counts and forms depending on the specific calligraphic style and context.
  • Annotation at stroke, character, word, and sentence levels provides fine-grained data crucial for advanced computational models in areas like online recognition and generative art.
  • The dataset successfully captures the artistic freedom and diversity inherent in various Arabic calligraphic styles, showcasing distinct visual forms for identical text.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.