shifted layout in scan recognition

Added by Simone Deparis over 9 years ago

Hello,

I have troubles in the detection of a page. It seems that the layout is shifted, although the corner circles seems to be correctly detected.

Did anyone have a similar problem? Is there something that I should try?

I have used several different DPI (from 200 to 600), grey scale and color.
I played a little bit with threasholds.
Always the same behavior.

The pages have been cut by 4mm before the scan, but this was also done in other occasions without troubles.

Any help is welcome.

Best regards
Simone Deparis


Replies (3)

RE: shifted layout in scan recognition - Added by Anirvan Sarkar over 9 years ago

Which version of AMC are you using? Also can you upload a sample tex file which illustrates this problem?

RE: shifted layout in scan recognition - Added by Simone Deparis over 9 years ago

version: 1.2.1463
OS: MacOSX (installed by port)

I am preparing a light version. Before that I try on a debian system to see if the problem persists.

RE: shifted layout in scan recognition - Added by Simone Deparis over 9 years ago

I found the problem!

In the preamble I was using the command
\geometry{a4paper,inner=2.7cm,outer=1.3cm,twoside,top=2.5cm}
to move the whole text a little on the side. Indeed we print a booklet on A3 paper which is stabled in the middle and, after correction, it is cut neatly.

On mac I did not get any warning during layout detection.
When checking on a debian system, I got a warning saying

"Some potential defects were detected for this subject. Correct them in the source and update the working documents.
The corner marks and binary boxes are not at the same location on all pages (See for exemple pages 1/1 and 1/2). "

I've set back the geometry to a centered one:

\geometry{a4paper,hmargin=2cm,top=2.5cm}

(note that the total width is exactly the same).
Everyting is now working as expected, even on MacOSX. I since the width is the same, I can scan the exam without issues.

Best regards
Simone

(1-3/3)