GMIC, great discovery

[geaplanet]

GMIC can apply filters to images really easily, and you can use it with a lot of files in a batch processing.
Here I used to clean the background without affecting letters.

[aku]

Are you talking about
http://gmic.eu/index.shtml
by chance ?

[geaplanet]

@aku, yes.

I used in the "repair" category, the "ians noise reduction" filter.
In that category, there is another filter called "repair scanned documents", it does a good job too for getting a total white background, but I don't like totally colors result.

Fot batch treatment, take the command from gimp/gmip filter (verbose to the layer name), in my case -iain_iains_nr 3,3,1,0,0,0,1,0,0,0,1.35,0,0, and then run in a terminal something like that:

gmic *.jpg -iain_iains_nr 3,3,1,0,0,0,1,0,0,0,1.35,0,0 -o img.jpg

And that's all, jpg files in a folder will be created with the filter applied.

[dingodog]

if you are running gmic in Linux, it would be very helpful to have gmic packaged as standalone package able to run (hopefully) on every linux distro, since compilation og gmic is not possible in older distro, while gmic is very useful

you need to download cde

http://www.pgbovine.net/cde.html

and package the gmic binary on your system with this syntax

Code: Select all

cde path...to/gmic

then you can upload the resulting gmic folder zipped somewhere

It would be very important and useful if you can do this attempt

[geaplanet]

@Dingodog, what system do you use? I'm running linux in an old Debian Wheezy, I think it is 2012 ¿? And do you have gmic even in Squeeze repository.

[dingodog]

first of all thanks for your reply (many thanks)

I'm currently running a distro (puppy linux 3.01) that is one of the few distro I found fully working on my outdated hardware (not to be snob

I seen the list of dependencies to build gmic, it is not so long but very heavy!

libgimp2.0-dev
libcv-dev
libhighgui-dev
libcurl4-openssl-dev
libfftw3-dev

I'm unable to compile these dependencies with gcc 4.1.2 and the current environment with glibc 2.5

so I kindly asked you make a portable gmic using cde; cde is not perfect but fails rarely and works the most part of times, what cde does is create a folder in directory you invoke, with executable (gmic in this case) with all dependencies it is able to track, mainly using ldd to identify these and isolating these in a layer that theorically makes an executable so packaged working in any linux environment independently from system libs

if you run cde to create a standalone gmic, any can test the package so made

[geaplanet]

I know about CDE and how it works, I talked to Philip Guo often years ago.
I'll try pack gmic with CDE, but if your machine is so old that you have to install puppylinux...processing images will be really heavy work for it. And don't forget "CDE packed apps" eat more RAM than normal compiled apps.

Only for curiosity, what machine do you have?

[dingodog]

I currently own a

PENTIUM IV, 2,6 Ghz, with 2 Gb of Ram (a 32 bit system)

the machine itself it not so old, but I had problems with wifi with other distro, since 7 years, instead, puppy linux perfectly fits my needs

I'm aware that cde is not perfect, but it is a test worth to do in my opinion (many thanks in advance)

the real problem is that development environment of Puppy Linux 3.01 comes with gcc 4.1.2 (but this is not the real problem) the real issue is the very old glibc (2.5) that I cannot upgrade without break the whole system, otherwise I had already compiled gmic by myself

P.S. I have several build of jbig2enc built with special parameters, and especially one with a patch that allows (with the help of PIL image library) to do what googlebooks does in their scanning work (having in the same page a b/w and a grayscale representation of page image for details uncodable in b/w) this patch by Alexey Kryukov also adds the ability to select the number of pages for dictionary for jbig2 encoder

Code: Select all

  -d --duplicate-line-removal: use TPGD in generic region coder
  -p --pdf: produce PDF ready data
  -P <number> --pages-per-dict <number>: pages per dictionary (default 15)
  -s --symbol-mode: use text region, not generic coder
  -t <threshold>: set classification threshold for symbol coder (def: 0.85)
  -T <bw threshold>: set 1 bpp threshold (def: 188)
  -r --refine: use refinement (requires -s: lossless)
  -O <outfile>: dump thresholded image as PNG
  -2: upsample 2x before thresholding
  -4: upsample 4x before thresholding
  -S: remove images from mixed input and save separately
  -j --jpeg-output: write images from mixed input as JPEG
  -v: be verbose

I'll upload a stand-alone portable as soon as possible

no time tonight to make this portable, but maybe it can be working also without to portabilize the executable, you can try my build of this special version of jbig2enc with patch by Alexey Kryukov

http://d-h.st/QSh7

it comes with a modified pdf.py script downloadable from my site
http://dokupuppylinux.info/_media/progr ... ica-pdf.py

[geaplanet]

Here you are: http://goo.gl/cDLTOQ

I don't know very well what you mean with that patch does, could you link an example?
I convert to pdf with pdfbeads, it chooses the best compressor for pages depending on their content.

[geaplanet]

For my old PC I use picaros, it has a lot of software installed and a lot of drivers (for wifi too). It is based on minino, another lightweight debian distribution.
It is especially for kids, but in menu -> minino utilities -> desktop style, you can choose a "normal" desktop based in lxde. Even it has a lxde tuned with gnome2 look.
All desktop only eat 100Mb RAM! And you can install all debian software.
http://minino.galpon.org/en/descargas