Search code examples
azure-cognitive-services

How to extract PRINTED and HANDWRITTEN text from a scanned document


Is it possible to have cognitive services return both the PRINTED and HANDWRITTEN text from a scanned document. I can OCR the document and regardless of whether I have TextRecognictionMode.Printed and TextRecognitionMode.Handwritten I get similar responses, but I was expecting to have a Json file returned with a property for the TYPE of text read as Printed or Handwritten as I'd like to identify Handwritten text after certain Printed text.

Is this possible at all?


Solution

  • I'm the Product Manager for OCR at Microsoft.

    This is currently not possible - as you noted, we have separate printed and handwriting OCR. That said, it is (firmly) on our roadmap - I cannot share the exact timeline, but we are working on it.

    Please stay tuned.