My team is currently experimenting with IXP’s capabilities for complex and unstructured documents after previously using the Document Understanding tool. For the most part, I’ve been impressed by its capabilities, but one particular issue I’ve noticed is some difficulty with extracting checkboxes and handwritten signatures. Most of the time, I can’t select these values on the page as a reference for the field. I believe this may be due to the OCR engine not recognizing them as text. I was wondering how other people had approached extracting these kinds of fields with IXP. Is there any type of change I can make to the OCR engine or base model that will help them be properly detected?
Hi Jarryd, it’s definitely helpful to know that signatures need to be listed as a text field in the taxonomy manager. I will keep that in mind when using IXP projects in studio. However, I don’t think that resolves the problems we’re having with checkboxes.
