Data capture mode

Added by Cedric Bernardin almost 10 years ago

For the data capture of my exam I used the "Some answers sheets were photocied" mode rather than the "Different answers sheets" mode. I think it was an error. I know that we can not switch from one mode to the other one. I produced an exam with several pages for each student (not fixed). Thus each exam sheet can contain two or three pages. All the sheets had a different identification. I don't really understand how the marks are obtained. Most of them seems to be correct but each exam sheet is splitted in one, two three pages (like 26:1 26:2 26:3 or 1:1). In front of each page, there is a mark (sometimes the same, sometimes different). Is there a solution to start a new data capture (for example by producing a new project for but by using the same exams sheets produced for the exam) ?


Replies (17)

RE: Data capture mode - Added by Alexis Bienvenüe almost 10 years ago

You can create a new project with the exact same source file, and re-run all the process (except printing !). It should work.

RE: Data capture mode - Added by Cedric Bernardin almost 10 years ago

Thanks. It works!

Cedric Bernardin wrote:

For the data capture of my exam I used the "Some answers sheets were photocied" mode rather than the "Different answers sheets" mode. I think it was an error. I know that we can not switch from one mode to the other one. I produced an exam with several pages for each student (not fixed). Thus each exam sheet can contain two or three pages. All the sheets had a different identification. I don't really understand how the marks are obtained. Most of them seems to be correct but each exam sheet is splitted in one, two three pages (like 26:1 26:2 26:3 or 1:1). In front of each page, there is a mark (sometimes the same, sometimes different). Is there a solution to start a new data capture (for example by producing a new project for but by using the same exams sheets produced for the exam) ?

RE: Data capture mode - Added by S K over 9 years ago

Hi.

I'm new to this forum, so apologies if my question has been answered before.
In short, I've installed AMC on Gentoo and was able to compile a quiz for 36 students,
with shuffling commendted out. Then I printed the questions from inside "okular",
along with the answer sheets. Then I put these separated answer sheets with students' ticked boxes
along with the boxes where they filled their student IDs (the IDs being of length 9, I changed the parameter accordingly).
Now, I scan the completed responses into a *tiff file, then try to recognise it in AMC,
but all I can get are the numbers "1/5/21" and so forth -- i.e. the numbers at the right-top -- but no data
on the boxes ticked, "--" for MSD and "sensitivity" fields.
What can be done, short of the manual input? Do you need more details?

Thank you.

RE: Data capture mode - Added by shreikant kv over 9 years ago

Pl try the following:

Open Doc-sujet and go to the bubble sheet of first copy/student in a pdf viewer.
Open the scan of bubble sheet of first copy/student in a pdf viewer (assuming you scanned to pdf).

Set the two at same magnification level. Compare whether they match in position.

Most likely they do not!

RE: Data capture mode - Added by S K over 9 years ago

Thank you for your reply. True, I've scanned into pdf, and the questions page and scans do not match when put at the same maginification level. How am I to change the scans to match?

RE: Data capture mode - Added by S K over 9 years ago

An update on my previous post: a multi-page pdf scan is split and converted by AMC to something with *pbm extension. And when I open these *pbms in okular and set same magnification level as that of the DOC-sujet.pdf, the positions, sure enough, do not match.

RE: Data capture mode - Added by S K over 9 years ago

One more update,

with "Different Answer Sheets" option I imported a multipage pdf file,
and the following output is found on the command-line:

Use of uninitialized value in string ne at /usr/lib/AMC/perl/AMC-gui.pl line 4573.
Use of the encoding pragma is deprecated at /usr/lib/AMC/perl/AMC-prepare.pl line 21.
ppmtoxpm: (Computing colormap...
ppmtoxpm: ...Done.  256 colors found.)
Use of uninitialized value in string ne at /usr/lib/AMC/perl/AMC-gui.pl line 4573.
Use of uninitialized value in string ne at /usr/lib/AMC/perl/AMC-gui.pl line 4573.
Can't call method "new" without a package or object reference at /usr/lib/AMC/perl/AMC-analyse.pl line 685, <GEN1> line 554.
Can't call method "new" without a package or object reference at /usr/lib/AMC/perl/AMC-analyse.pl line 685, <GEN1> line 611.
Can't call method "new" without a package or object reference at /usr/lib/AMC/perl/AMC-analyse.pl line 685, <GEN1> line 1394.
Can't call method "new" without a package or object reference at /usr/lib/AMC/perl/AMC-analyse.pl line 685, <GEN1> line 1393.
Can't call method "new" without a package or object reference at /usr/lib/AMC/perl/AMC-analyse.pl line 685, <GEN1> line 552.
[...]
Use of uninitialized value in numeric gt (>) at /usr/lib/AMC/perl/AMC-gui.pl line 5136.
Use of uninitialized value in string ne at /usr/lib/AMC/perl/AMC-gui.pl line 4573.
Use of uninitialized value in string ne at /usr/lib/AMC/perl/AMC-gui.pl line 4573.
ppmtoxpm: (Computing colormap...
ppmtoxpm: ...Done.  256 colors found.)
ppmtoxpm: (Computing colormap...
ppmtoxpm: ...Done.  256 colors found.)
ppmtoxpm: (Computing colormap...
ppmtoxpm: ...Done.  256 colors found.)
ppmtoxpm: (Computing colormap...
ppmtoxpm: ...Done.  256 colors found.)
ppmtoxpm: (Computing colormap...
ppmtoxpm: ...Done.  256 colors found.)
ppmtoxpm: (Computing colormap...
ppmtoxpm: ...Done.  256 colors found.)

So, some modules are missing? I'm using GentooLinux.

"Command output details" box contains the following:

Page : +30/1/33+
Page : +29/3/34+
Page : +29/2/35+
Page : +29/1/36+
Page : +24/1/51+
Page : +30/3/31+
[...]

Indeed, a quiz consists of three pages. The result: no scans recognized. Oddly enough, when I choose "Some papers were photocopied", the scans do appear alongside the pages for manual entry.

RE: Data capture mode - Added by shreikant kv over 9 years ago

One source of this mismatch:
The printer zoomed the page and/or centered/adjusted margins etc from DOC-sujet before printing onto paper.
Another source of this mismatch:
The scanner did some zooming/centering.

Pl look at scanner settings and ensure that scanner is not the culprit. If scanner is the only culprit, it
can be fixed easily -- just rescan with proper settings.

If printer is the culprit: Have a look at graphicx and geometry packages in LaTeX.

You can include one scanned page (no margins, no parindent, no parkip : geometry package is useful here) at
different scaling levels and experiment. This will be a bit painful -- but not impossible. Suppose through
hit and miss you realize that a scaling factor is .8 results in perfect recognition. Then include all
pdfpages at this scaling factor and hopefull things will work.

RE: Data capture mode - Added by S K over 9 years ago

Hm, now I rescanned everything, opened a new project, and in the "cr" directory I can find perfectly recognized pages such as the one attached.
However, the AMC-GUI shows nothing (attached picture). what am I to do?

page-1-3.jpg (580.5 kB)

AMCtroubleshooting.png (43.2 kB)

RE: Data capture mode - Added by shreikant kv over 9 years ago

Most likely related to "incorrect/incomplete" installation. Pl try to get
access to another machine say with ubuntu +install AMC+run your project
there...

RE: Data capture mode - Added by Alexis Bienvenüe over 9 years ago

Or try to run AMC in debug mode (from Help menu) and have a look at the produced log file to see where is the problem.

RE: Data capture mode - Added by S K over 9 years ago

Short of a fresh installation, I'm running it in the background. The log-file is getting bigger and bigger. Where am I to submit it, when the automatic capture is complete?
As to the ubuntu+AMC mix, I'm sure I've tried it before Gentoo, but there were some problems with Perl modules, so that the GUI's buttons were inactive. The current Gentoo installation works with manual input.

RE: Data capture mode - Added by S K over 9 years ago

Ubuntu "sudo apt-get" works fine, except that when launched from a commandline, AMC yields
Gtk-WARNING **: Unable to locate theme engine in module_path: "pixmap", at /usr/lib/perl5/Gtk2.pm line 138.
Can't locate XML/LibXML/SAX.pm in @INC (@INC contains: /etc/perl /usr/local/lib/perl/5.14.2 /usr/local/share/perl/5.14.2 /usr/lib/perl5 /usr/share/perl5 /usr/lib/perl/5.14 /usr/share/perl/5.14 /usr/local/lib/site_perl .) at (eval 19) line 1.

RE: Data capture mode - Added by S K over 9 years ago

There is no an "edit" option for one's own posts, so my apologies for pursuing this, but for future sufferers for Ubuntu:
I got around the previous problem by issuing "sudo apt-get install libxml-libxml-perl-*"
Now I launch "auto-multiple-choice" and I get the following error in the console:
Unable to locate theme engine in module_path: "pixmap", at /usr/lib/perl5/Gtk2.pm line 138
plus the buttons of GUI are all inactive.
Please transfer this to the appropriate thread if this is off-topic, and thanks for your replies.
I've not yet given up on making AMC OCR work on Gentoo.

RE: Data capture mode - Added by Alexis Bienvenüe over 9 years ago

plus the buttons of GUI are all inactive.

Maybe you are confused with the Unity layout from Ubuntu, which moves the menus to the top left of the whole screen (not attached to the window).

RE: Data capture mode - Added by S K over 9 years ago

OK, Finally it works with Ubuntu! Even the OCR.
Thank you all.

RE: Data capture mode - Added by S K over 9 years ago

Sorry, can't help adding: the manual association feature with the students' names in boxes is simply fantastic...

(1-17/17)