CoolFace
Datasetpublic

DimV-Ai/kuzushiji-character-dataset-ogihan-v1

Kuzushiji Character Dataset (Ogihan / Ogi Domain) This dataset contains single-character Kuzushiji image crops derived from the Ogihan (小城藩) historical materials, published in a format compatible with datasets released by CODH (Center for Open Data in the Humanities). The dataset is designed for: Kuzushiji OCR Character-level recognition Multimodal and vision–language model training Comparative research with existing CODH datasets Dataset Structure The dataset… See the full description on the dataset page: https://huggingface.co/datasets/DimV-Ai/kuzushiji-character-dataset-ogihan-v1.

sourceHugging Faceupdated 9mo agoView on Hugging Face
1likes13downloads
Dataset Card

Kuzushiji Character Dataset (Ogihan / Ogi Domain)

This dataset contains single-character Kuzushiji image crops derived from the Ogihan (小城藩) historical materials, published in a format compatible with datasets released by CODH (Center for Open Data in the Humanities).

The dataset is designed for:

  • —Kuzushiji OCR
  • —Character-level recognition
  • —Multimodal and vision–language model training
  • —Comparative research with existing CODH datasets

Dataset Structure

The dataset is organized per manuscript (book), following CODH conventions.

Image Data

  • —Character images are stored inside tar.gz archives in the tars/ directory.
  • —Each image is a single character crop extracted from page images.
  • —Image filenames encode Unicode, manuscript ID, page number, and coordinates.

Filename format:

Example:

Metadata (Coordinate CSV)

Each manuscript has a corresponding coordinate file in metadata/.

CSV Columns

Column NameDescription
UnicodeUnicode code point of the character (e.g. U+4E03)
ImagePage image identifier (<book_id>_<page_id>)
XX coordinate (top-left) of the bounding box
YY coordinate (top-left) of the bounding box
WidthWidth of the bounding box
HeightHeight of the bounding box
Block IDEmpty (not available for Ogihan data)
Char IDAnnotation identifier from the original data

Coordinates are in pixel units, relative to the original page image.

Manuscripts Included

  • —N0100501901
  • —N0100502001

Each manuscript is processed independently, with page numbering restarting from 0001 per book.

Data Creation Notes

  • —Bounding boxes and character annotations were created by human annotators.
  • —Character crops were generated from page images using these annotations.
  • —The dataset was converted into CODH-compatible format for interoperability.

License and Attribution

Please cite appropriately when using this dataset. Licensing follows the terms of the original Ogihan materials.

Related Datasets

  • —Combined Ogihan + CODH dataset: https://huggingface.co/datasets/DimV-Ai/kuzushiji-character-dataset-v1
  • —CODH Kuzushiji resources: https://codh.rois.ac.jp/char-shape/book/