5 JSON files containing OCR-extracted Arabic text from book pages. Each entry maps a pageId to a source image url, a pdfPageNumber, and the corresponding ocrOutput string.
Use Cases
- Train OCR models for Arabic script by pairing the url image assets with the ocrOutput text
- Develop text-to-image alignment algorithms using the ocrOutput and pageId identifiers
- Perform linguistic analysis on historical Arabic legal texts found within the ocrOutput field
Strengths
- 5 JSON files containing structured page-level OCR data
- Direct url links to PNG image assets for every OCR entry
- ocrOutput field containing Arabic text with specific section markers like '١٠٢٧ - مسألة'
- pdfPageNumber field for tracking the original position within the source document