Is there a way to increase the accuracy of the OCR in DTPO?
In preferences, is 150 dpi and 75% enough - will I get better accuracy if I raise it to 300 dpi and 100% accuracy?
Would I get increased accuracy using some other OCR engine outside of DTPO and then importing the searchable PDF into the DTPO database? If so, what’s the most accurate engine?
And finally … I noticed that the latest version of DTPO is using the Abbyy FineReader 8.x engine. Would I benefit using the Abbyy FineReader 11 engine outside DTPO? Or is there any way to get Abbyy 11 into DTPO?
We recommend that the scanner resolution should be 300 dpi for a good good image of text, which makes recognition easier and more accurate when OCR is performed on the image.
The DTPO Preferences > OCR settings for dpi and image quality are for the image saved after OCR, so does not affect accuracy of text recognition.
I don’t see significant differences in OCR accuracy between the OCR module in the current version of DTPO and ABBY FineReader Pro 12.x. Using on both of the same scanner images, I’ve seen a few differences, especially in recognition of small print or unusual fonts, but overall about the same very good accuracy percentage. Version 8 got some text right that version 12 Pro got wrong, and vice versa.
I agree that there aren’t accuracy difference between the OCR done by DTPO or the ABBY program.
However, the size of the files that each produces is definitely different. I don’t consider myself an expert, so YMMV, but I consistently found that the ABBY program provided a significantly smaller file size. So for me the ABBY program is my go to for OCR work rather than DTPO.