Scan Tailor

[Tulon]

boost/make_shared.hpp is part of boost, and appeared in boost version 1.39, which is relatively new. I'll probably replace it with a more portable construct, but meanwhile you will have to update your boost installation,

[laszlo]

Thank you for your help. (libboost1.37 is enough)

I compiled it, and I still cant find where is this option hided...
Is it available in the gui? If I open an existing project can I see it, or should I create a new one?

[Tulon]

It is available in the GUI (a checkbox "Dewarp") on the Output stage, but has serious limitations:
1. Only works in B/W mode.
2. Expects output resolution of 600 DPI (the default one).
3. Only works in simple cases.

[laszlo]

It is available in the GUI (a checkbox "Dewarp") on the Output stage, but has serious limitations:
1. Only works in B/W mode.
2. Expects output resolution of 600 DPI (the default one).
3. Only works in simple cases.

I dont have this option:-(

I recompiled it, but still no luck. Could you please test on your side? (fresh git checkout, and compile).
I think something is still missing from the source.

[Tulon]

Sorry, I forgot to commit some of the stuff. Try now.

[laszlo]

I grepped in the source for "Dewarp", but only found this one:
filters/output/BlackWhiteOptions.h:48

Code: Select all

void setDewarp(bool enabled) { m_dewarp = enabled; }


I think something is still missing. Could you please clarify where should this
"Dewarp" gui code (the checkbox) located in the sourcecode?

Best regards,
Laszlo

[laszlo]

Sorry, I forgot to commit some of the stuff. Try now.

Ahh ok. I pulled, Im compiling now. Will report in ten minutes.

[laszlo]

I tried it out, but it was rather a disappointment...

I thought my page is only a little warped, but the "dewarp" feature could not fix it, even made it worse.

Here is a picture without dewarp:


Here is a picture with dewarp activated:


And here is the original photo for reference. Hope it helps something.

[rob]

Apparently it only works in very simple cases

[laszlo]

From an amateurish point of view, I would think the best method is to build the 3d model of a deformed page.
Then apply the inverse to get a plan.

I would also think, that a page can only deform in one direction. So a half-sphere is not possible.
Only the plan deformed in one direction.