OCR - does the engine autodetect rotation of images in the PDF?

[osos]

I am trying to optimize OCR scanning using the library.

It seems that rotated images with text are not recognized or purely recognized.

How can I ensure that rotation of particular pages are identified and prerotated before the OCR engine scans the page?

Thanks in advance

[Surya Kumar]

Hi Johannes, 

 

Thank you for using Syncfusion products. 

 

We do have support for auto detecting the rotated images in the PDF page and perform OCR,  please enable the “AutoDetectRotation” property in OCRProcessor settings. Please find the code snippet for the same below: 

            using (OCRProcessor processor = new OCRProcessor(@"../../TesseractBinaries/")) 

 

            { 

 

                //Load a PDF document 

 

                PdfLoadedDocument lDoc = new PdfLoadedDocument(DataPathBase + "RotatedOCR1.pdf"); 

                 

                processor.Settings.Language = "eng"; 

                //Enabling AutoDetectRotation property 

                processor.Settings.AutoDetectRotation = true; 

                processor.PerformOCR(lDoc, @"../../Tessdata/"); 

                

                //Saving and closing the document 

                lDoc.Save(DataPathOutput + "Output1.pdf"); 

                lDoc.Close(true); 

 

                 

 

 

            } 

 

 

Also, please follow below instruction:  

“Tessdata” folder should contain the file “osd.traineddata” in order to get this feature working which can be downloaded from below link: 

http://www.syncfusion.com/downloads/support/directtrac/156562/ze/osd-2095924609  

We have created sample for the same which can be found in below link: 

http://www.syncfusion.com/downloads/support/forum/132660/ze/Sample_OCR764172180 

Please let us know if you need any further information 

Regards, 

Surya Kumar