RECEIPT OCR REVIEW TEMPLATE
JPG to Excel team | 2026-09-30

Purpose: compare receipt OCR outputs against the same manually checked sources.
This is a blank review worksheet, not benchmark results or a tax-compliance asset.

1. TEST SET AND REFERENCE
Source image ID:
Merchant / language / currency:
Image type: screenshot / camera photo / scan
Condition: clear / faded / wrinkled / glare / cropped / other
Reviewer's name or initials:
Manually checked reference location:
Reference rules for dates, separators and identifiers:
Sensitive details removed before upload:

2. EXPECTED RECEIPT FIELDS
Merchant as printed:
Transaction date as printed:
Normalized date (only if unambiguous):
Receipt number:
Currency:
Subtotal:
Discounts / refunds:
Tax:
Tip:
Final total:
Amount paid / cash tendered:
Change:
Expected line-item count:

3. TOOL RUN
Tool / model / version (if documented):
Date of run:
Mode: text / spreadsheet / structured fields
Input settings:
Output format:
Usage charge or credits:
Upload time:
Recognition time:
Correction time:
Export time:

4. FIELD AND ROW CHECKS
Field | reference | OCR result | exact match? | correction / uncertainty
Merchant:
Date:
Receipt number:
Subtotal:
Discount:
Tax:
Tip:
Final total:
Currency:

Line # | description | quantity | unit price | discount | line total | issue

Missing line items:
Duplicate line items:
Wrapped descriptions placed in the wrong row:
Uncertain characters or amounts (do not guess):
Identifier or date auto-formatting changes in Excel:

5. RECONCILIATION
Visible line-item sum:
Discounts / refunds:
Tax / tip adjustments following this receipt's labels:
Calculated expected total:
Printed final total:
Difference:
Mismatch explained and reviewed by:

6. SUMMARY
Correct required fields / all required fields:
Correct required rows / all required rows:
Total time = upload + recognition + correction + export:
Total cost = service charge + correction labor + setup/maintenance allocation:
Usable checked result? yes / no
Reason for choosing or rejecting this tool:
Original image and reviewed output kept together at:

Repeat the same reference, settings and evaluation rules for each comparison tool.
Dataset results are not a substitute for a representative test on your receipts.