Japanese Synthetic OCR 150K: Font Images for Text Recognition

Japanese Synthetic OCR 150K is a dataset of synthetic images for optical character recognition tasks. The dataset was created by the author 'deepcopy' and was last updated on July 7, 2025. Its source is listed as a Kaggle dataset containing Japanese font images.

Use Cases

• Train optical character recognition models based on synthetic Japanese font images.
• Benchmark OCR model performance on Japanese characters.
• Augment training data for document analysis systems with synthetic text.
• Study the robustness of text recognition algorithms to different font styles.

Strengths

• Dataset title suggests a scale of approximately 150,000 items.
• Last update timestamp of 2025-07-07 09:11:27 is provided.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

https://www.kaggle.com/datasets/hnthnt/jp-font-image-dataset-02

Freshness

Last updated 2025-07-07 09:11:27; freshness should be verified.