Translate Cloud OCR parameters to Fine Reader command line parameters

  • 112 Views
  • Last Post 15 March 2017
kofoapp posted this 13 March 2017

Hi, we have been using the cloud ocr for quite some time now and are very happy with the results. Due to the high number of processing tasks that our application now has on a daily basis, we are in the process of switching to a server version of the Fine Reader 11 engine. Now our problem is that we can't reproduce the exact results we are getting from the Cloud OCR, as there are so many different command line parameters. The question: Would it be possible to specify the exact request parameters we send to the cloud OCR and get the exact command line parameters used internally? Thanks a lot!

Order By: Standard | Newest | Votes
Oksana Serdyuk posted this 14 March 2017

Could you please specify your Application ID and what recognition settings you use and want to translate to FREngine?

kofoapp posted this 15 March 2017

Hi Oksana, thanks for the quick reply. This is the request we send to the cloud OCR: http://cloud.ocrsdk.com/processImage?language=german,english&imageSource=photo&correctOrientation=true&correctSkew=true&profile=textExtraction&xml:writeRecognitionVariants=true&exportFormat=pdfSearchable,xmlForCorrectedImage Is there a way I can send you our application id by pm or email?

Oksana Serdyuk posted this 15 March 2017

In order to replicate these settings in FineReader Engine 11 you need to set the TextExtraction_Accuracy predefined profile using the LoadPredefinedProfile method of the Engine object and then set the following additional settings :

[RecognizerParams]
TextLanguage = German,English

[PagePreprocessingParams]
GeometryCorrectionMode = GCM_Correct
CorrectOrientation = TRUE
CorrectSkew = TSPV_Auto

[PDFExportParams]
TextExportMode = PEM_ImageOnText

[XMLExportParams]
WriteNondeskewedCoordinates = FALSE
WriteCharacterRecognitionVariants = TRUE

kofoapp posted this 15 March 2017

great, thank you! we'll try to implement that

Close