OCR recognition problems


#1

The OCR function in Architect does not consistently or accurately recognize the characters from pdf files scanned using the software. It cannot distinguish background light images, tilted objects, it confuses M for N and similar issues. It is also difficult to select characters read by the software to cut and paste to other documents.


#2

Hi,

which version of PDF Architect are you currently using?
OCR isn’t always able to produce 100% accurate results, but we can check if there is anything we can improve on our end.
Are you scanning with the highest quality settings?

Best regards

Robin


#3

They are using the most current version and yes they did set the settings to be the best.
This is what it looks like:

This is what it SHOULD look like:


#4

Hi,

it does indeed look quite unusable; did you set the OCR recognition language to “English”?
The setting “all languages” seems to often generate bad results.
The background noise in the scan could however also cause some issues, as it is full of visible dots.

Best regards

Robin


#5

Thank you -
Yes they have tried all the settings and have it set to English.
We most likely will have this department switch to Foxit.