Organisation with a large paper archive
AI Document Scanning System
A system that reads scanned documents and saves the data automatically. More than 2 million images processed.
The client's name is kept private at their request.
- images processed
- 2M+
- processing, no manual typing
- Batch
The problem
The client had a huge archive of paper records. Staff were typing details from each document into a database by hand.
It was slow and costly, and errors were common.
What I built
I built a document scanning system that reads each scanned image with OCR, picks out the important fields and saves them as structured data in a database.
It runs in batches, handles poor-quality scans, and flags documents that need a person to check them.
Technology
- Python
- Tesseract OCR
- Database