Japanese And Korean Handwriting OCR Samples From 100 Subjects

Comprising handwriting OCR data from 100 subjects, including 50 Japanese, 49 Korean, and 1 Afghan individual. It was collected using multiple cellphone models and features different text corpora per subject.

Use Cases

• Train an OCR model to recognize Japanese and Korean handwriting from images of 100 subjects.
• Analyze handwriting style variations across the 50 Japanese and 49 Korean subject cohorts.
• Benchmark OCR performance on data collected from multiple cellphone models.

Strengths

• Data from 100 distinct subjects provides a sample of individual handwriting styles.
• Includes subjects from two primary nationalities: 50 Japanese and 49 Korean.
• Collection involved multiple cellphone models, introducing device diversity.

Limitations

• The dataset is described as a sample, suggesting it is not a full-scale collection.
• Limited demographic representation with only one subject from Afghanistan.
• Specific details on image count, resolution, or annotation quality are unavailable.

Provenance

Source

Nexdata

Collection Method

Collected from 100 subjects handwriting on different cellphone models with varying text corpora.

Freshness

Last updated on 2025-04-24.

Geography

Primarily Japan and Korea, based on subject nationality.