Hi,
We bought cloud OCR SDK. during our use, we found this software does not highlight not-sure items. instead the PC version highlighted the not-sure items in yellow.
Can you advise how we can solve this?
Thanks,
Dale
Cloud OCR SDK does not highlight not-sure items
Was this article helpful?
0 out of 0 found this helpful
Comments
1 comment
HI Dale!
As far as I understand you would like to have additional information about uncertainly recognized characters. For this purpose you could export recognition result to an XML file by setting the exportFormat parameter to "xml" when calling the Cloud OCR API processing methods (processImage, processDocument etc.) The structure of the resulting XML files is described here: Output XML Document.
For your purposes, you could check whether the suspicious attribute of the charParams tags is set to true (suspicious="1") or to false (suspicious is omitted). If its value is true ("1"), the corresponding character is uncertainly recognized. You could also set the xml:writeRecognitionVariants parameter of the processing method to true to get additional information such as character recognition confidence (charConfidence attribute of the charRecVariant tag). This way you will be able to set your own threshold value and only apply manual correction only when character recognition confidence is below this value.
After getting the information about which characters are suspicious, you can get the characters regions and post-process your document by highlighting these characters.
Please sign in to leave a comment.