OCR issues
Improve receipt image quality and understand what incomplete extraction results usually mean.
OCR problems usually start with the image, not with the reward logic. When OmniLab cannot read the merchant, amount, or other key fields reliably, the receipt is more likely to become INCOMPLETE, INVALID, or harder to validate confidently.
The product's own photo guidance
The receipt challenge already asks participants for an image that is:
- clear
- sharp
- centered on the receipt
- showing the merchant and amount visibly
Reinforce the same guidance in your campaign copy and staff training.
Typical OCR problems
| Problem | Likely result |
|---|---|
| Blurry or cropped photo | Missing merchant or amount |
| Folded or shadowed receipt | Incomplete extraction |
| Low contrast thermal print | Weak OCR confidence |
| Same image uploaded repeatedly | Duplicate-image errors or duplicate receipt status |
What to do when OCR keeps failing
- test the challenge with several real receipt formats, not just one ideal sample
- simplify the participant instructions around receipt capture
- use manual validation if field variability is high and reward risk justifies it
- review whether the confidence posture or field requirements are stricter than the campaign needs
Important workflow rule
Only INCOMPLETE receipts are meant for participant resubmission. If the issue is recurring, do not ask users to guess endlessly. Improve the image instructions or review the receipt manually.