AI
Arabic ID OCR Pipeline
Reads scanned Egyptian national ID cards in Arabic, validates every field and stores clean records
- Year
- 2024
- Status
- Shipped
- Role
- OCR & data-extraction module — graduation project team
- Stack
- Python · Tesseract OCR · OpenCV · MySQL
Problem
My graduation project, MAVS, was a multi-modal authentication and verification system. It needed reliable personal data from scanned Egyptian national ID cards: Arabic script, varied scan quality, and no room for a wrong field ending up in the database.
Approach
I built the OCR and data-extraction module end to end: image pre-processing, Arabic-script field recognition, parsing into structured records, rule-based validation, and persistence to MySQL.
Result
Extracted fields were tested against the source images, and failure cases such as low-quality scans and partially readable fields were handled before anything reached storage.
