Manga109-s: 87 Japanese Manga Volumes for Commercial Use

The Manga109-s dataset is a subset of the Manga109 dataset containing 87 Japanese manga books available for commercial use. It is intended for academic and non-academic applications, including image analysis and object detection tasks. The dataset is hosted by the University of Tokyo.

Use Cases

• Perform object detection on manga images to identify and localize characters and objects.
• Train character recognition models on Japanese text and artwork found within the 87 manga volumes.
• Analyze visual styles and layouts across the commercially available subset of the Manga109 collection for computer vision research.

Strengths

• Contains 87 distinct manga books, providing a defined corpus for analysis.
• Specifically curated for commercial use, offering clear licensing terms for a subset of the original Manga109 data.
• Hosted and maintained by an academic institution (University of Tokyo), suggesting structured curation.

Limitations

• The specific size, format, and annotation details (e.g., bounding boxes, text) for the 87 volumes are not provided in the input.
• The dataset scope is limited to 87 works, which may not represent the full diversity of manga genres or publishers.
• Sample data and column structure are unavailable, making it difficult to assess the data's readiness for specific tasks.

Provenance

Source

hal-utokyo (University of Tokyo), derived from the Manga109 dataset.

Freshness

The dataset was last updated on 2025-05-23.

Geography

Japan, based on the content and associated tags.