Page 1 of 1

Preventing Page 2 from Showing through Page 1

PostPosted: Wed Sep 29, 2004 4:12 am
by mgilvey
Have you ever had to scan printed material and after the scan appears you see a ghost image of the text that was on the opposite side of the page. How do you fix it? You could clip the white point untill the ghost text disappears but if there is a photo on the scanned side of the page you might clip some of the 1/4 tone by doing this unless you took the time to make a mask to protect the photo area.

What has worked for me is to place a sheet of Black (not white) paper behind the page you are scanning. This makes the black text on the opposite page blend into the black. True, it may darken the white a little but if you set the white point after putting the black paper behind, it should come up nicely with very little (if any) remnants of text from the other page showing up.

PostPosted: Wed Sep 29, 2004 8:13 am
by LSI_Noack
Dear mgilvey

yes, I had the same positive results with a sheet of black cardboard material. Thanks for sharing this insight with the other forum users.

Personally, I like to use greyscale rather than lineart when scanning text for OCR. Then I apply the white and black level with SilverFast's pipette tool which gives me a quick result which works well enough with OCR programmes. Though this might sound like a little extra amount of time, it is actually faster doing it this way as there are less corrections to be made in the OCR text results later-on.

Kind regards
Sonny Noack
- tech support, LSI -

PostPosted: Wed Sep 29, 2004 2:50 pm
by mgilvey
That's awesome Sonny, thanks!!!

Re:

PostPosted: Mon Dec 14, 2015 10:27 am
by tonytong
LSI_Noack wrote:Dear mgilvey

yes, I had the same positive results with a sheet of black cardboard material. Thanks for sharing this insight with the other forum users.

Personally, I like to use greyscale rather than lineart when scanning text for ocr. Then I apply the white and black level with SilverFast's pipette tool which gives me a quick result which works well enough with OCR programmes. Though this might sound like a little extra amount of time, it is actually faster doing it this way as there are less corrections to be made in the OCR text results later-on.

Kind regards
Sonny Noack
- tech support, LSI -


appreciate, useful information.