Annotated dataset of 30 receipt images with bounding boxes for structured field extraction (OCR training).
Merchant Name | Date | Total | Subtotal | Tax | Payment Method | Address
- 180 fields annotated (≈6 per receipt)
- 100% accurracy on 20-sample audit
- Edge cases handled (low quality, missing fields, format variation)
Exports/receipt_extraction_export.json— bounding box annotations (JSON)docs— labeling guidelines + project summaryscreenshots— annotation proof
Bounding box annotation • OCR data prep • multi-field extraction • handling messy real-world data