We have a process that involves reading from PDF documents which are scanned images of handwritten customer forms. Is it a good candidate to automate? The UiPath version we will be using is 2018.4.6.
We checked the PDF documents - we noticed the handwritten text includes tick marks, crosses, strike-through text etc. So can such requirements be considered for automation? Kindly let me know your valuable thoughts?
Thanks in advance!
Handwritten PDF automation would be relaying on the OCR tool we gonna use.
We tried with ABBY few months back, but was not able to achieve expected results. Like we tried a complex solution of IPA. Fusing RPA with ML.
But market ready OCR tool like ABBY, Microsoft and Google OCRs are good to convert hand written notes to text.
Thanks @adheedhan for your valuable inputs! But will such automation yield good accuracy rate? I know it depends on the legibility of the text. But can OCRs extract the text exactly like a human vision? The concerning point is if we mark such activities feasible and finally get to know in later stages of development it extracts correct text only 10 % of the time, likewise.
Yes the accuracy was good with the ABBY which i was using. You can use FineReader from ABBY which can extract the text alone from the documents.
Try using UiPath In-built OCRs before you reach for ABBY, so you dont need extra licenses to be acquired.
I think ABBYY is not free, please correct me if wrong.
In built OCRs like Google and Tesseact is not helping.
Yes ABBY comes with a paid license.
But you can get a trail version to test it.