One of my clients requires high success rates on OCR of text that is often times in a halftone region. So far, I have not had any success with Acrobat 9 Pro Text Recognition over any degree of halftone. Can this process work under any circumstances on text over halftones?
I've spent several hours testing and reading what little information is available on this issue, based on the lack of complaints it seems no one else is having this issue.
Resolution: 400, 600
Color: B/W (JBIG2, CCIT G4)
Scanner: Fujitsu fi-5750C
Thanks in advance!
"halftone" implies the actual zone of pixels at halftone will have a horizontal/vertical effective resolution less than that of the scan setting. Less "line weight"/"line density", no?
Use Acrobat's Preflights. There are pre-configured profiles available for assessing resolution(s).
If effective resolution is low, contrast low, weight low => OCR can't "grab" stuff in such a region.
For ClearScan, a bitmap image of the character is left. So, not enough to 'capture' (including enough to designate as a 'suspect') then goes to the image - To edit, would have to go after it (these left behind bitmap images) with the photo editor.
Be well...
Be well...