My organisation has a form that we issue out to our customer’s. Customers complete the form with ink and return it. We are planning to scan these forms on return and extract the customers email address and phone number using UiPath Intelligent Document Processing, updating our internal system where a change in contact information is identified.
A small proof of concept has found that the UiPath OCR Engine is not achiving the required accuracy in order to push these changes. For Example, we have achieved an 82% character accuracy on the email field, however only 21% of email addresses were 100% accurate.
Im looking for some tips and suggestions on how I can improve the accuracy of the OCR Engine on handwritten fields please.
You cannot increase the accuracy of a specific OCR engine, the OCR engine is the engine and cannot be ‘tuned’ by the end user in this case.
Your options are limited to changing the OCR engine, there are others on the market, some specifically designed for hand writing, that might have better success rates. You should benchmark against some different OCR engines to see what works.
