Is it possible to train the OCR assertiveness?


#1

Dear Supporters,

Is it possible to train the OCR assertiveness for the Portuguese language (PT-BR)?
There is already a package of Portuguese language, but is it possible to improve it?

What are the minimum requirements for good assertiveness?
I am using Aspose.OCR, the version: 17.6.0.0 .net C#
Resource: Portuguese language resource file for Aspose.OCR for .NET 3.2.0

Regards,
Tiago.


#2

@tiago.silva

Thanks for contacting support.

Yes, Aspose.OCR supports recognizing Portuguese Language. Support Fonts, Image formats and Styles are as following:

  • Support for JPG, JPEG, PNG, GIF, BMP and TIFF image file formats for OCR.
  • Read popular fonts including Arial, Times New Roman, Courier New, Verdana, Tahoma and Calibri.
  • Support for regular, bold and italic font styles.

Furthermore, the minimum DPI of an image should be 300 in order to get expected results. Font size is also a factor which should be considered while using the API. Font with larger size results in accurate recognition of characters. In case you need further information, please feel free to let us know.