From djmuhlestein at gmail.com  Tue Jan  2 21:42:39 2007
From: djmuhlestein at gmail.com (Dennis Muhlestein)
Date: Tue, 2 Jan 2007 13:42:39 -0700
Subject: [Image-SIG] PNG Support with Parsed images.
Message-ID: <eeed88270701021242j6d138334oceb1d0d96b9a719e@mail.gmail.com>

If I load a png file with Image.open, I can manipulate it, save it
etc.  If I however use ImageFile.Parser(), I can read the size, format
and any info, but I can't show or manipulate the image in any way.

I get this exception on any of the requested methods:

Here is what I'm typing:

>>> import ImageFile
>>> f=open('somefile.png')
>>> p=ImageFile.Parser()
>>> p.feed(f.read())
>>> i=p.close()
>>> i.show()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/lib/python2.4/site-packages/PIL/Image.py", line 1449, in show
    _showxv(self, title, command)
  File "/usr/lib/python2.4/site-packages/PIL/Image.py", line 2082, in _showxv
    file = image._dump(format=format)
  File "/usr/lib/python2.4/site-packages/PIL/Image.py", line 476, in _dump
    self.load()
  File "/usr/lib/python2.4/site-packages/PIL/ImageFile.py", line 189, in load
    s = read(self.decodermaxblock)
  File "/usr/lib/python2.4/site-packages/PIL/PngImagePlugin.py", line
365, in load_read
    return self.fp.read(bytes)
  File "/usr/lib/python2.4/site-packages/PIL/ImageFile.py", line 300, in read
    data = self.data[pos:pos+bytes]
TypeError: unsubscriptable object


This is with version 1.1.6.  I double checked 1.1.5 and got the same thing.

This works fine with a JPEG image.  The PNG support appears to have a
bug to me, but perhaps I'm missing something.

Thanks
Dennis

From marshallpenguin at gmail.com  Tue Jan  2 22:02:09 2007
From: marshallpenguin at gmail.com (Joey Marshall)
Date: Tue, 2 Jan 2007 18:02:09 -0300
Subject: [Image-SIG] Colorize an image
Message-ID: <73cf534f0701021302o18b3cb44m80a507825d09b983@mail.gmail.com>

Hello,
I am having a difficulty with developing my game. I would like to
colorize the outfits on my penguins
(http://joey101.net/projects/snowballz/). I have them separated out to
make it easier to colorize... but I haven't been able to figure out
how to.

I already have PIL as a dependency for snowballz so I would like to
use it, but there is nothing in PIL to do this easily. Does anyone
know how to colorize an image? (all images are (RGBA) in red and I
would like to be able to use a color like: (0,0,255) to colorize it
to)

Any help would be awesome.

Thanks!
Joey

-- 
http://joey101.net - my blog on web development

From tinodb at gmail.com  Tue Jan  2 22:16:00 2007
From: tinodb at gmail.com (TiNo)
Date: Tue, 2 Jan 2007 16:16:00 -0500
Subject: [Image-SIG] Compiling PIL with JPEG support does not work at all
In-Reply-To: <435b46e50612271018hb77475eoe1a118ed59777f7f@mail.gmail.com>
References: <435b46e50612271018hb77475eoe1a118ed59777f7f@mail.gmail.com>
Message-ID: <435b46e50701021316y61932e2fv2f7d5f04352f4a0c@mail.gmail.com>

Nobody any idea?

2006/12/27, TiNo <tinodb at gmail.com>:
> Hi,
>
> I know there are many people experiencing the same problem, and there
> are different solutions around, but I am still unable to compile PIL
> with JPEG support.
>
> I am on a shared webhost (bluehost) and have build python 2.5. I am
> now trying to build PIL.
>
> I have build jpeg-6b from http://www.ijg.org/  in
> /home/spafenet/lib/lib/jpeg (that now contains bin/ and man/)
>
> My setup.py looks like this (relevant part:):
> ---------------------------------------------------------------------
> # --------------------------------------------------------------------
> # Library pointers.
> #
> # Use None to look for the libraries in well-known library locations.
> # Use a string to specify a single directory, for both the library and
> # the include files.  Use a tuple to specify separate directories:
> # (libpath, includepath).  Examples:
> #
> # JPEG_ROOT = "/home/libraries/jpeg-6b"
> # TIFF_ROOT = "/opt/tiff/lib", "/opt/tiff/include"
> #
> # If you have "lib" and "include" directories under a common parent,
> # you can use the "libinclude" helper:
> #
> # TIFF_ROOT = libinclude("/opt/tiff")
>
> FREETYPE_ROOT = None
> JPEG_ROOT = "/home/spafenet/lib/lib/jpeg/bin"
> TIFF_ROOT = None
> ZLIB_ROOT = "/home/spafenet/lib/lib/zlib/lib","/home/spafenet/lib/lib/zlib/include"
> TCL_ROOT = None
>
> ------------------------------------------------------------------------------------
>
> I have also tried JPEG_ROOT = "/home/spafenet/lib/lib/jpeg/" but that
> doesn't work either.
>
> Building report (for both options):
>
> --------------------------------------------------------------------
> PIL 1.1.6 BUILD SUMMARY
> --------------------------------------------------------------------
> version       1.1.6
> platform      linux2 2.5 (r25:51908, Dec 18 2006, 13:56:40)
>               [GCC 3.4.6 20060404 (Red Hat 3.4.6-3)]
> --------------------------------------------------------------------
> *** TKINTER support not available
> *** JPEG support not available
> --- ZLIB (PNG/ZIP) support ok
> *** FREETYPE2 support not available
> --------------------------------------------------------------------
> To add a missing option, make sure you have the required
> library, and set the corresponding ROOT variable in the
> setup.py script.
>
> To check the build, run the selftest.py script.
>
> ----------------------------------------------------------------------------
>
> It works with the ZLIB though. As it takes my 1.2.3 version over the
> preinstalled 1.2.1.2 version. (got a warning previously). But it
> refuses to accept the JPEG.
>
> Any suggestions?
>

From fredrik at pythonware.com  Wed Jan  3 10:02:58 2007
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 3 Jan 2007 10:02:58 +0100
Subject: [Image-SIG] Compiling PIL with JPEG support does not work at all
References: <435b46e50612271018hb77475eoe1a118ed59777f7f@mail.gmail.com>
Message-ID: <enfrfl$bub$1@sea.gmane.org>

"TiNo" <tinodb at gmail.com> wrote:

> I know there are many people experiencing the same problem, and there
> are different solutions around, but I am still unable to compile PIL
> with JPEG support.
>
> I am on a shared webhost (bluehost) and have build python 2.5. I am
> now trying to build PIL.
>
> I have build jpeg-6b from http://www.ijg.org/  in
> /home/spafenet/lib/lib/jpeg (that now contains bin/ and man/)

PIL needs the library and include files (*.a/*.so and *.h), not the binaries.

</F> 




From newz at bearfruit.org  Wed Jan  3 00:52:01 2007
From: newz at bearfruit.org (Matthew Nuzum)
Date: Tue, 2 Jan 2007 17:52:01 -0600
Subject: [Image-SIG] Colorize an image
In-Reply-To: <73cf534f0701021302o18b3cb44m80a507825d09b983@mail.gmail.com>
References: <73cf534f0701021302o18b3cb44m80a507825d09b983@mail.gmail.com>
Message-ID: <f3c0b4080701021552o726235bbt9b8517e0833f2b93@mail.gmail.com>

On 1/2/07, Joey Marshall <marshallpenguin at gmail.com> wrote:
>
> Hello,
> I am having a difficulty with developing my game. I would like to
> colorize the outfits on my penguins
> (http://joey101.net/projects/snowballz/). I have them separated out to
> make it easier to colorize... but I haven't been able to figure out
> how to.
>
> I already have PIL as a dependency for snowballz so I would like to
> use it, but there is nothing in PIL to do this easily. Does anyone
> know how to colorize an image? (all images are (RGBA) in red and I
> would like to be able to use a color like: (0,0,255) to colorize it
> to)
>
>
You will likely get much better answers from others on the list, but I've
done it once. I experimented with a few different techniques and I did it
simply as an experiment. I took my images in photoshop, made them black and
white (er, actually, black and transparent) by grabbing the channel that had
the appearance and deleting all the other channels. I then saved it out as a
png with alpha transparency. That's a quick overview of how I got my png,
use whatever technique works for you.

Once I had my image, I loaded it into PIL and figured out the dimensions. I
then created a new image and set its color to red or orange or whatever I
wanted. Then I overlaid my image with alpha transparency onto the new image.

I can't remember exactly what I did and I'm looking for the source and can't
find it easily. However, the concept is pretty simple and you likely already
know how to do all of that stuff anyway.
-- 
Matthew Nuzum
www.bearfruit.org
newz2000 on freenode
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070102/6b0b8ebc/attachment.htm 

From fredrik at pythonware.com  Thu Jan  4 20:41:22 2007
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Thu, 04 Jan 2007 20:41:22 +0100
Subject: [Image-SIG] Error with last version PIL Image (AttributeError:
	pixel_access)
In-Reply-To: <e6bc38c10612201255s716e7b7dgc6eb91895a19e0cb@mail.gmail.com>
References: <e6bc38c10612201255s716e7b7dgc6eb91895a19e0cb@mail.gmail.com>
Message-ID: <enjl92$4an$1@sea.gmane.org>

barchia wrote:

> Today I update the PIL, because  I will try the new version for
> check if a problem with the font are solved, instead i found a
> new problem as i write in object...
> 
> this is the trace of error:
> 
>  >>> import Image, ImageDraw
>  >>> im = Image.new("RGB",(30,30),'white')
>  >>> draw = ImageDraw.Draw(im)
> Traceback (most recent call last):
>   File "<interactive input>", line 1, in ?
>   File "C:\Python24\Lib\site-packages
> \PIL\ImageDraw.py", line 294, in Draw
>     return ImageDraw(im, mode)
>   File "C:\Python24\Lib\site-packages\PIL\ImageDraw.py", line 59, in 
> __init__
>     im.load()
>   File "C:\Python24\Lib\site-packages\PIL\Image.py", line 599, in load
>     return self.im.pixel_access(self.readonly)
> AttributeError: pixel_access

if you installed PIL 1.1.6 over an earlier version, make sure that it's 
not picking an outdated _imaging.pyd file.  type

     >>> import Image
     >>> Image.core

to see what pyd file it's using.

</F>


From david.giesen at kodak.com  Thu Jan  4 22:00:00 2007
From: david.giesen at kodak.com (david.giesen at kodak.com)
Date: Thu, 4 Jan 2007 16:00:00 -0500
Subject: [Image-SIG] Capturing the Tkinter Canvas to the Windows Clipboard
Message-ID: <OF46959B55.F8B18A43-ON85257259.0073033C-85257259.00735B8B@knotes.kodak.com>

Hi -

I've searched all over and I can't figure out how to transfer an image of 
the Tkinter Canvas to the Windows clipboard (or any clipboard, for that 
matter).  I've use the PIL to capture and display an image of the Canvas, 
but I can't figure out how to put that image into the clipboard directly. 
Yes, I can capture it, display it automatically using Windows Paint and 
use "copy" in that program to put it in the clipboard, but that seems, 
well, clunky.

Do you have any suggestions how I might do this?  I've spent a 
considerable amount of time googling this to no avail, so I'm sorry if 
this is an FAQ that I should know.

Thanks in advance for any help!

Dave Giesen


From Chris.Barker at noaa.gov  Thu Jan  4 21:16:44 2007
From: Chris.Barker at noaa.gov (Christopher Barker)
Date: Thu, 04 Jan 2007 12:16:44 -0800
Subject: [Image-SIG] Fw: Python - WEIRD Code needed (help with)
In-Reply-To: <028701c72659$bc02bf70$0a03010a@mail2world.com>
References: <028701c72659$bc02bf70$0a03010a@mail2world.com>
Message-ID: <459D60AC.6040802@noaa.gov>

You'll need to be a little more clear about what you're asking:

1) how to call a python program from VB -- I have no idea, try 
comp.lang.python, or a PyWin32 mailing list if there is one. Although 
calling a python program as a command line program from VB should be 
trivial -- but why use Python for that? Why not go all the way?

2) How to do this all in Python -- If you want a GUI, you need to choose 
a toolkit first -- I recommend wxPython, but TkInter, PYGTK, PyQT, etc 
are good options, too.

3) How to do the image generation/manipulation -- then PIL (and this 
list) may be the best place, but you could also just do it straight with 
the GUI toolkit (wxPython anyway). You might also want to use the 
"numpy" package -- it's a good way to manipulate arrays of numbers.

4) If you want the equivalent of video feedback, but without actual 
images, just numbers, then numpy is definitely a good way to go.

If I wanted to write a program that demonstrated video feedback, I'd 
probably do the number crunching with numpy, and make a GUI to display 
and manipulate the results in wxPython. I'm not sure where VB would fit 
into this.


-- 
Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R            (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chris.Barker at noaa.gov

From fredrik at pythonware.com  Fri Jan  5 08:06:32 2007
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 05 Jan 2007 08:06:32 +0100
Subject: [Image-SIG] Capturing the Tkinter Canvas to the Windows
	Clipboard
In-Reply-To: <OF46959B55.F8B18A43-ON85257259.0073033C-85257259.00735B8B@knotes.kodak.com>
References: <OF46959B55.F8B18A43-ON85257259.0073033C-85257259.00735B8B@knotes.kodak.com>
Message-ID: <enktdo$2hg$1@sea.gmane.org>

david.giesen at kodak.com wrote:

> I've searched all over and I can't figure out how to transfer an image of 
> the Tkinter Canvas to the Windows clipboard (or any clipboard, for that 
> matter).  I've use the PIL to capture and display an image of the Canvas, 
> but I can't figure out how to put that image into the clipboard directly. 
> Yes, I can capture it, display it automatically using Windows Paint and 
> use "copy" in that program to put it in the clipboard, but that seems, 
> well, clunky.

PIL can copy from the clipboard, but not to it.  here's the relevant 
issue tracker item:

     http://effbot.python-hosting.com/ticket/23

contributions are welcome!  (a ctypes-based implementation would be good 
start, I think).

</F>


From colin.barnette at gmail.com  Fri Jan  5 09:44:49 2007
From: colin.barnette at gmail.com (Colin Barnette)
Date: Fri, 5 Jan 2007 03:44:49 -0500
Subject: [Image-SIG] Image.load() issues.
Message-ID: <aac92b50701050044q5a4e9fcfh5cde2bb9bff32060@mail.gmail.com>

I'm having some really odd trouble with PIL. As shown on in the Handbook I have:

 import Image

 print Image.VERSION
 img = Image.load("sushi.jpg")

  And the output I get is this:

 1.1.6
 Traceback (most recent call last):
   File "pichtml.py", line 5, in ?
     img = Image.load("sushi.jpg")
 AttributeError: 'module' object has no attribute 'load'

  So I tried:

 img = Image.Image.load("sushi.jpg")

  Traceback (most recent call last):
   File "pichtml.py", line 3, in ?
     img = Image.Image.load("sushi.jpg")
 TypeError: unbound method load() must be called with Image instance
as first argument (got str instance instead)

  And finally:

 img = Image.Image()
 img.load("sushi.jpg")

  Traceback (most recent call last):
   File "D:\pichtml.py", line 4, in ?
     img.load("sushi.jpg")
 TypeError: load() takes exactly 1 argument (2 given)

  What am I doing wrong?

 Any help or ideas would be appreciated.

From fredrik at pythonware.com  Fri Jan  5 13:27:37 2007
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 5 Jan 2007 13:27:37 +0100
Subject: [Image-SIG] Image.load() issues.
References: <aac92b50701050044q5a4e9fcfh5cde2bb9bff32060@mail.gmail.com>
Message-ID: <enlg7p$c2n$1@sea.gmane.org>

Colin Barnette wrote:

> I'm having some really odd trouble with PIL. As shown on in the Handbook I have:

what handbook is this?

> import Image
>
> print Image.VERSION
> img = Image.load("sushi.jpg")

if you want to open an image file, use "open":

    img = Image.open("sushi.jpg")

</F> 




From david.giesen at kodak.com  Fri Jan  5 14:06:59 2007
From: david.giesen at kodak.com (david.giesen at kodak.com)
Date: Fri, 5 Jan 2007 08:06:59 -0500
Subject: [Image-SIG] Capturing the Tkinter Canvas to the Windows Clipboard
Message-ID: <OF6BCEC9FC.423383F4-ON8525725A.0047DE84-8525725A.00480D22@knotes.kodak.com>

OK, thanks for the info.  Unfortunately, programming such a thing is far 
beyond my skills, so I won't be able to contribute much.

I take it there is probably no non-PIL-based solution either?

Thanks again -

Dave Giesen

From sittner at lkb.ens.fr  Fri Jan  5 15:24:00 2007
From: sittner at lkb.ens.fr (sittner at lkb.ens.fr)
Date: Fri, 5 Jan 2007 15:24:00 +0100 (CET)
Subject: [Image-SIG] installation problem for imaging 1.1.6 on ubuntu 6.06
In-Reply-To: <mailman.4.1167994803.19625.image-sig@python.org>
References: <mailman.4.1167994803.19625.image-sig@python.org>
Message-ID: <55228.129.199.118.95.1168007040.squirrel@mailgate.phys.ens.fr>

Hello everyone and happy new year,
i am trying to install imaging 1.1.5 on my ubuntu 6.06 machine, and i get
this message:

r:/usr/lib/python2.4/site-packages/Imaging-1.1.5$ python setup.py install
running install
running build
running build_py
running build_ext
building '_imaging' extension
gcc -pthread -fno-strict-aliasing -DNDEBUG -g -O2 -Wall
-Wstrict-prototypes -fPIC -IlibImaging -I/usr/include -I/usr/local/include
-I/usr/include/python2.4 -c _imaging.c -o
build/temp.linux-i686-2.4/_imaging.o
In file included from
/usr/lib/gcc/i486-linux-gnu/4.1.2/include/syslimits.h:7,
                 from /usr/lib/gcc/i486-linux-gnu/4.1.2/include/limits.h:11,
                 from /usr/include/python2.4/Python.h:18,
                 from _imaging.c:74:
/usr/lib/gcc/i486-linux-gnu/4.1.2/include/limits.h:122:61: error:
limits.h: No such file or directory
In file included from _imaging.c:74:
/usr/include/python2.4/Python.h:32:19: error: stdio.h: No such file or
directory
/usr/include/python2.4/Python.h:34:5: error: #error "Python.h requires
that stdio.h define NULL."
/usr/include/python2.4/Python.h:37:20: error: string.h: No such file or
directory
/usr/include/python2.4/Python.h:38:19: error: errno.h: No such file or
directory
.
.
.
.
/usr/include/python2.4/pyport.h:73: error: expected ?=?, ?,?, ?;?, ?asm?
or ?__attribute__? before ?Py_uintptr_t?
/usr/include/python2.4/pyport.h:74: error: expected ?=?, ?,?, ?;?, ?asm?
or ?__attribute__? before ?Py_intptr_t?
/usr/include/python2.4/pyport.h:94:76: error: math.h: No such file or
directory
/usr/include/python2.4/pyport.h:101:22: error: sys/time.h: No such file or
directory
/usr/include/python2.4/pyport.h:102:18: error: time.h: No such file or
directory
/usr/include/python2.4/pyport.h:120:24: error: sys/select.h: No such file
or directory
/usr/include/python2.4/pyport.h:159:22: error: sys/stat.h: No such file or
directory
.
.
.
.
_imaging.c: In function ?_getcodecstatus?:
_imaging.c:2574: warning: return from incompatible pointer type
_imaging.c: In function ?_save_ppm?:
_imaging.c:2608: warning: return from incompatible pointer type
_imaging.c:2611: warning: return from incompatible pointer type
_imaging.c: At top level:
_imaging.c:2709: error: initializer element is not constant
_imaging.c:2709: error: (near initialization for ?methods[53].ml_name?)
_imaging.c:2709: error: initializer element is not constant
_imaging.c:2709: error: (near initialization for ?methods[53].ml_meth?)
_imaging.c: In function ?_getattr?:
_imaging.c:2733: warning: passing argument 3 of
?PyCObject_FromVoidPtrAndDesc? from incompatible pointer type
_imaging.c:2735: warning: return from incompatible pointer type
_imaging.c: At top level:
_imaging.c:2778: error: initializer element is not constant
_imaging.c:2778: error: (near initialization for ?Imaging_Type.ob_type?)
_imaging.c:2799: error: initializer element is not constant
_imaging.c:2799: error: (near initialization for ?ImagingFont_Type.ob_type?)
_imaging.c:2811: error: initializer element is not constant
_imaging.c:2811: error: (near initialization for ?ImagingDraw_Type.ob_type?)
_imaging.c:2966: error: initializer element is not constant
_imaging.c:2966: error: (near initialization for ?functions[37].ml_name?)
_imaging.c:2966: error: initializer element is not constant
_imaging.c:2966: error: (near initialization for ?functions[37].ml_meth?)
error: command 'gcc' failed with exit status 1

what's the problem here?


From michele.petrazzo at unipex.it  Fri Jan  5 17:28:09 2007
From: michele.petrazzo at unipex.it (Michele Petrazzo)
Date: Fri, 05 Jan 2007 17:28:09 +0100
Subject: [Image-SIG] installation problem for imaging 1.1.6 on ubuntu
 6.06
In-Reply-To: <55228.129.199.118.95.1168007040.squirrel@mailgate.phys.ens.fr>
References: <mailman.4.1167994803.19625.image-sig@python.org>
	<55228.129.199.118.95.1168007040.squirrel@mailgate.phys.ens.fr>
Message-ID: <459E7C99.1020101@unipex.it>

sittner at lkb.ens.fr wrote:
> Hello everyone and happy new year,

Also to you :)

> i am trying to install imaging 1.1.5 on my ubuntu 6.06 machine, and i
> get this message:
> 

A couple of questions:
- why the 1.1.5? There is the 1.1.6
- why by sources? ubuntu hasn't the pil package from its repository?
   on Debian:
   michele:~$ apt-cache search python imaging
   <-cut->
   python-imaging
- have you installed the libc6 (I think, or libc5) -dev package?
   always on debian:
   michele:~$ dpkg -L libc6-dev | grep limits
   /usr/include/limits.h
   michele:~$

So, reading your error, I see a lot of "No such file or directory", so
install the python-dev, libc6-dev and so on...
Or (better solution), apt-get install python-imaging!

Bye,
Michele

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3273 bytes
Desc: S/MIME Cryptographic Signature
Url : http://mail.python.org/pipermail/image-sig/attachments/20070105/b30e97d9/attachment.bin 

From Chris.Barker at noaa.gov  Fri Jan  5 22:41:15 2007
From: Chris.Barker at noaa.gov (Christopher Barker)
Date: Fri, 05 Jan 2007 13:41:15 -0800
Subject: [Image-SIG] Capturing the Tkinter Canvas to the Windows
	Clipboard
In-Reply-To: <OF6BCEC9FC.423383F4-ON8525725A.0047DE84-8525725A.00480D22@knotes.kodak.com>
References: <OF6BCEC9FC.423383F4-ON8525725A.0047DE84-8525725A.00480D22@knotes.kodak.com>
Message-ID: <459EC5FB.4000704@noaa.gov>

david.giesen at kodak.com wrote:

> I take it there is probably no non-PIL-based solution either?

wxPython can take a PIL image, and can put images on the Clipboard (I'm 
pretty sure -- I've never done it) . Maybe tk can too.

-Chris




-- 
Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R            (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chris.Barker at noaa.gov

From marshallpenguin at gmail.com  Sat Jan  6 16:25:21 2007
From: marshallpenguin at gmail.com (Joey Marshall)
Date: Sat, 6 Jan 2007 12:25:21 -0300
Subject: [Image-SIG] Colorize an image
In-Reply-To: <f3c0b4080701021552o726235bbt9b8517e0833f2b93@mail.gmail.com>
References: <73cf534f0701021302o18b3cb44m80a507825d09b983@mail.gmail.com>
	<f3c0b4080701021552o726235bbt9b8517e0833f2b93@mail.gmail.com>
Message-ID: <73cf534f0701060725v58e930e0h3d38e2f7d3909a04@mail.gmail.com>

On 1/2/07, Matthew Nuzum <newz at bearfruit.org> wrote:
> On 1/2/07, Joey Marshall <marshallpenguin at gmail.com> wrote:
> > Hello,
> > I am having a difficulty with developing my game. I would like to
> > colorize the outfits on my penguins
> > (http://joey101.net/projects/snowballz/). I have them
> separated out to
> > make it easier to colorize... but I haven't been able to figure out
> > how to.
> >
> > I already have PIL as a dependency for snowballz so I would like to
> > use it, but there is nothing in PIL to do this easily. Does anyone
> > know how to colorize an image? (all images are (RGBA) in red and I
> > would like to be able to use a color like: (0,0,255) to colorize it
> > to)
> >
> >
>
> You will likely get much better answers from others on the list, but I've
> done it once. I experimented with a few different techniques and I did it
> simply as an experiment. I took my images in photoshop, made them black and
> white (er, actually, black and transparent) by grabbing the channel that had
> the appearance and deleting all the other channels. I then saved it out as a
> png with alpha transparency. That's a quick overview of how I got my png,
> use whatever technique works for you.
>
> Once I had my image, I loaded it into PIL and figured out the dimensions. I
> then created a new image and set its color to red or orange or whatever I
> wanted. Then I overlaid my image with alpha transparency onto the new image.
>
> I can't remember exactly what I did and I'm looking for the source and can't
> find it easily. However, the concept is pretty simple and you likely already
> know how to do all of that stuff anyway.
> --
> Matthew Nuzum
> www.bearfruit.org
> newz2000 on freenode


That would be cool if you could find the source. I must admit that I'm
very new to PIL and image manipulation itself. Thanks Matthew.

If anyone else has help on this please do tell! :)

-- 
http://joey101.net - my blog on web development

From fredrik at pythonware.com  Sat Jan  6 16:33:18 2007
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sat, 06 Jan 2007 16:33:18 +0100
Subject: [Image-SIG] installation problem for imaging 1.1.6 on ubuntu
	6.06
In-Reply-To: <55228.129.199.118.95.1168007040.squirrel@mailgate.phys.ens.fr>
References: <mailman.4.1167994803.19625.image-sig@python.org>
	<55228.129.199.118.95.1168007040.squirrel@mailgate.phys.ens.fr>
Message-ID: <enoffu$u2i$1@sea.gmane.org>

sittner at lkb.ens.fr wrote:

> i am trying to install imaging 1.1.5 on my ubuntu 6.06 machine, and i get
> this message:

> /usr/lib/gcc/i486-linux-gnu/4.1.2/include/limits.h:122:61: error:
> limits.h: No such file or directory

your C compiler installation is not complete.

     $ sudo apt-get install build-essential

should do the trick.

</F>


From nicolas at famillepinault.fr  Tue Jan  9 22:09:01 2007
From: nicolas at famillepinault.fr (Nicolas Pinault)
Date: Tue, 09 Jan 2007 22:09:01 +0100
Subject: [Image-SIG] RGBA to P problem
Message-ID: <45A4046D.9040505@famillepinault.fr>

Hi,

I have an application where I need to convert RGBA images(png) to P 
images (gif or png).
I use im.convert("P") to do the conversion. The problem comes with alpha 
channel. The resulting P mode image does not contain any tranparency 
information. Transparency is set to none and transparent pixels are set 
to black (in my test).
I tried to add an alpha channel after the conversion with 
im.putalpha(alpha). This converts the image back to RGBA mode.
Is there a smart way to do the convertion ?

Nicolas Pinault

From ned at nedbatchelder.com  Tue Jan  9 22:40:53 2007
From: ned at nedbatchelder.com (Ned Batchelder)
Date: Tue, 09 Jan 2007 16:40:53 -0500
Subject: [Image-SIG] Clipped text, still.
Message-ID: <45A40BE5.7010904@nedbatchelder.com>

I tried installing 1.1.6, mostly to see the improvements in clipped 
text.   PIL 1.1.5 would clip pieces of text that extended beyond the 
character width, and would wrap those pixels around to the other side of 
the image.  So "joe" might lose part of the bulb of the j and display it 
below the e.

But 1.1.6 still clips the bulb.  It no longer transports the lost pixels 
to the other end of the string, but the clipping still happens.  Is 
there some new calling technique that I don't know about to get all of 
the text displayed?  I also see low descenders still being clipped (like 
the bottoms of g's).

I'm running on Windows, Python 2.4, PIL 1.1.6 installed from binaries, 
using Georgia font.

--Ned.

-- 
Ned Batchelder, http://nedbatchelder.com

From Scott.Daniels at Acm.Org  Wed Jan 10 07:09:25 2007
From: Scott.Daniels at Acm.Org (Scott David Daniels)
Date: Tue, 09 Jan 2007 22:09:25 -0800
Subject: [Image-SIG] Clipped text, still.
In-Reply-To: <45A40BE5.7010904@nedbatchelder.com>
References: <45A40BE5.7010904@nedbatchelder.com>
Message-ID: <eo1vsd$a2n$1@sea.gmane.org>

Ned Batchelder wrote:
> I tried installing 1.1.6, mostly to see the improvements in clipped 
> text.   PIL 1.1.5 would clip pieces of text that extended beyond the 
> character width, and would wrap those pixels around to the other side of 
> the image.  So "joe" might lose part of the bulb of the j and display it 
> below the e.
> 
> But 1.1.6 still clips the bulb.  It no longer transports the lost pixels 
> to the other end of the string, but the clipping still happens.  Is 
> there some new calling technique that I don't know about to get all of 
> the text displayed?  I also see low descenders still being clipped (like 
> the bottoms of g's).

Sounds like you are running into the edge of the writing area,
but expect PIL to ignore the clipping boundary for letters that
are partially inside the clipped area.

-- Scott David Daniels
Scott.Daniels at Acm.Org


From lab_monkey at mail2world.com  Sat Jan  6 06:39:29 2007
From: lab_monkey at mail2world.com (Ior Bock)
Date: Fri, 5 Jan 2007 21:39:29 -0800
Subject: [Image-SIG] Video Feedback Emulation
Message-ID: <2f5a601c73155$08e56120$70cb010a@mail2world.com>


Thanks Chris,

That helps steer me in the right direction for tool selection. What Im
specifically stuck on however, is a basic idea on how
to do emulate video feedback.....

<-----Original Message----->

From: Christopher Barker [Chris.Barker at noaa.gov]
Sent: 1/4/2007 12:16:44 PM
To: Ior Bock; image-sig at python.org
Subject: Re: [Image-SIG] Fw: Python - WEIRD Code needed (help with)

You'll need to be a little more clear about what you're asking:

1) how to call a python program from VB -- I have no idea, try 
comp.lang.python, or a PyWin32 mailing list if there is one. Although 
calling a python program as a command line program from VB should be 
trivial -- but why use Python for that? Why not go all the way?

2) How to do this all in Python -- If you want a GUI, you need to choose
a toolkit first -- I recommend wxPython, but TkInter, PYGTK, PyQT, etc 
are good options, too.

3) How to do the image generation/manipulation -- then PIL (and this 
list) may be the best place, but you could also just do it straight with
the GUI toolkit (wxPython anyway). You might also want to use the 
"numpy" package -- it's a good way to manipulate arrays of numbers.

4) If you want the equivalent of video feedback, but without actual 
images, just numbers, then numpy is definitely a good way to go.

If I wanted to write a program that demonstrated video feedback, I'd 
probably do the number crunching with numpy, and make a GUI to display 
and manipulate the results in wxPython. I'm not sure where VB would fit 
into this.


-- 
Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception

Chris.Barker at noaa.gov
. 


Click to consolidate your debt and lower your monthly expenses
<http://tagline.bidsystem.com/fc/CAaCDCZ6sVDYMd0ujETmdCts0BYMOuT2/> 



<span id=m2wTl><p><font face="Arial, Helvetica, sans-serif" size="2" style="font-size:13.5px">_______________________________________________________________<BR>Get the Free email that has everyone talking at <a href=http://www.mail2world.com target=new>http://www.mail2world.com</a><br>  <font color=#999999>Unlimited Email Storage &#150; POP3 &#150; Calendar &#150; SMS &#150; Translator &#150; Much More!</font></font></span>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070105/e3d2c9bf/attachment.htm 

From dshimer at gmail.com  Thu Jan 11 10:42:23 2007
From: dshimer at gmail.com (Dennis Shimer)
Date: Thu, 11 Jan 2007 04:42:23 -0500
Subject: [Image-SIG] Tiff not reading dpi
Message-ID: <99ea8b7b0701110142g5e59ffa0t687eaaf71051b844@mail.gmail.com>

I have a tif image that at least 3 different programs can see the
correct dpi information in
(Irfanview, Gimp, Alchemy) but when I read the image using pil 1.6

print im.info  -  shows only {'compression': 'raw'}
dpix,dpiy=im.info['dpi']  -  gives an exception

and if I set it explicitely then using
im.save(filename,dpi=(dpix,dpiy))

Irfanview sees the new dpi.
Gimp complains that "resoultion units are meaningless"
and Alchemy sees no DPI info at all but doesn't complain.

Is there any reason that there is so much inconsistancy in results, or
a way to write dpi with pil that would be universally recognized?

From dshimer at gmail.com  Thu Jan 11 12:42:03 2007
From: dshimer at gmail.com (Dennis Shimer)
Date: Thu, 11 Jan 2007 06:42:03 -0500
Subject: [Image-SIG] Copying Tiff tags to new image
Message-ID: <99ea8b7b0701110342w14e2c7bdx5a256e79071f8c83@mail.gmail.com>

I am doing some basic modification to existing TIFF files using PIL
(1.6 in python 2.5). When I open the image there is quite a bit of
info in the tags dict that I would like to save with the new image.
Apparently (based on other forums I've seen) once you process the
image all the TIFF specific data gets lost. I can copy the info but
can't find an example of how to copy or re-assign it to the new image.
For example

import Image,TiffImagePlugin
im=Image.open('oldname.tif','r')
tags=im.tag.tags
im.rotate(90,0).save('newname.tif')

will open an image named oldname.tif, save the tags, then save a new
image that has been rotated by 90. How can I assign those tags back to
the new image? Or what important concepts am I just totally missing?

From Travis.Cochran at analog.com  Thu Jan 11 15:53:30 2007
From: Travis.Cochran at analog.com (Cochran, Travis)
Date: Thu, 11 Jan 2007 09:53:30 -0500
Subject: [Image-SIG] PIL 1.1.6 eval function
Message-ID: <4F19D8A3D7B26247A4481AE28348A848049FE0C3@gboexm2.ad.analog.com>

I'm trying to use the Image.eval function on a JPEG image. I can't seem
to find the right syntax for entering the function and it gives me
errors for the variables I try to use, "R,G,B". Any idea what I'm doing
wrong?
-Travis Cochran

From dshimer at gmail.com  Thu Jan 11 16:44:00 2007
From: dshimer at gmail.com (Dennis Shimer)
Date: Thu, 11 Jan 2007 10:44:00 -0500
Subject: [Image-SIG] PIL 1.1.6 eval function
In-Reply-To: <99ea8b7b0701110743l744bfb4cyf086bf313a1b8cee@mail.gmail.com>
References: <4F19D8A3D7B26247A4481AE28348A848049FE0C3@gboexm2.ad.analog.com>
	<99ea8b7b0701110743l744bfb4cyf086bf313a1b8cee@mail.gmail.com>
Message-ID: <99ea8b7b0701110744l2e881938nd74fda2f27b93b5b@mail.gmail.com>

This worked for me.

open an image object, I'll call im

Write a function that you would like applied to each pixel of im, here
is a dumb one that will just double the value this making the new
image very very bright.

def function(pixel):
    return pixel*2

Then eval the image based on the function

Image.eval(im2,function).show()

or save it to a new image which you can save

newimage=Image.eval(im2,function)

Dennis Shimer


On 1/11/07, Cochran, Travis <Travis.Cochran at analog.com> wrote:
> I'm trying to use the Image.eval function on a JPEG image. I can't seem
> to find the right syntax for entering the function and it gives me
> errors for the variables I try to use, "R,G,B". Any idea what I'm doing
> wrong?
> -Travis Cochran
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>

From robin at squiredeck.com.au  Sat Jan 13 11:52:39 2007
From: robin at squiredeck.com.au (Robin Whitlock)
Date: Sat, 13 Jan 2007 18:52:39 +0800
Subject: [Image-SIG] Plone Imaging User
Message-ID: <20070113104735.M38521@squiredeck.com.au>

Hi,
I felt I would ilke to comment on why it was so hard to get PIL installed into my Plone
2.5 environment on an X86_64 FC3 box. This is not a winge, but more like it would be
nice to have a 64 bit howto in the docs somewhere so the less priveliged may find the
install easier.

As it happens I needed to go back to 1.1.4 so that I could edit the Make file, is this
still not an option with the later versions (to change compiler options).

Hope this can lead to helping others, thanks for your time.

--
Robin Whitlock
Director - Squiredeck

C/- Open WebMail Project (http://openwebmail.org)


From dshimer at gmail.com  Thu Jan 11 17:12:53 2007
From: dshimer at gmail.com (Dennis Shimer)
Date: Thu, 11 Jan 2007 11:12:53 -0500
Subject: [Image-SIG] PIL 1.1.6 eval function
In-Reply-To: <4F19D8A3D7B26247A4481AE28348A848049FE1ED@gboexm2.ad.analog.com>
References: <4F19D8A3D7B26247A4481AE28348A848049FE1ED@gboexm2.ad.analog.com>
Message-ID: <99ea8b7b0701110812n594dbc34i30e2bda7e1d5bdf5@mail.gmail.com>

How about if you
im=Image.open('red.jpg','r')

which will give you open jpeg object then

Image.eval(im,function).show()

or if you want to do it all on one line

Image.eval(Image.open('red.jpg','r'),function).show()

Dennis

On 1/11/07, Cochran, Travis <Travis.Cochran at analog.com> wrote:
> I tried  that and got this error.
>
>
> Traceback (most recent call last):
>   File "C:\Documents and Settings\tcochra\My Documents\test", line 6, in
> -toplevel-
>     Image.eval("red.jpg",function).show()
>   File "C:\Python24\Lib\site-packages\PIL\Image.py", line 1979, in eval
>     return image.point(args[0])
> AttributeError: 'str' object has no attribute 'point'
>
>
>
> Could it be a problem with the IJG JEPG library? I reinstalled PIL and
> still get the error. Thanks for you help.
>
> -Travis Cochran
>
> -----Original Message-----
> From: Dennis Shimer [mailto:dshimer at gmail.com]
> Sent: Thursday, January 11, 2007 10:43 AM
> To: Cochran, Travis
> Subject: Re: [Image-SIG] PIL 1.1.6 eval function
>
> This worked for me.
>
> open an image object, I'll call im
>
> Write a function that you would like applied to each pixel of im, here
> is a dumb one that will just double the value this making the new image
> very very bright.
>
> def function(pixel):
>     return pixel*2
>
> Then eval the image based on the function
>
> Image.eval(im2,function).show()
>
> or save it to a new image which you can save
>
> newimage=Image.eval(im2,function)
>
> Dennis Shimer
>
>
> On 1/11/07, Cochran, Travis <Travis.Cochran at analog.com> wrote:
> > I'm trying to use the Image.eval function on a JPEG image. I can't
> > seem to find the right syntax for entering the function and it gives
> > me errors for the variables I try to use, "R,G,B". Any idea what I'm
> > doing wrong?
> > -Travis Cochran
> > _______________________________________________
> > Image-SIG maillist  -  Image-SIG at python.org
> > http://mail.python.org/mailman/listinfo/image-sig
> >
>

From ned at nedbatchelder.com  Wed Jan 10 19:12:19 2007
From: ned at nedbatchelder.com (Ned Batchelder)
Date: Wed, 10 Jan 2007 13:12:19 -0500
Subject: [Image-SIG] Clipped text, still.
In-Reply-To: <eo1vsd$a2n$1@sea.gmane.org>
References: <45A40BE5.7010904@nedbatchelder.com> <eo1vsd$a2n$1@sea.gmane.org>
Message-ID: <45A52C83.2070102@nedbatchelder.com>

In this case, there is no clipping boundary, and there are plenty of 
pixels to accommodate the extremeties of the letters.  The wrapping has 
been fixed, but the clipping still occurs.  Is this true or not?

--Ned.

Scott David Daniels wrote:
> Ned Batchelder wrote:
>   
>> I tried installing 1.1.6, mostly to see the improvements in clipped 
>> text.   PIL 1.1.5 would clip pieces of text that extended beyond the 
>> character width, and would wrap those pixels around to the other side of 
>> the image.  So "joe" might lose part of the bulb of the j and display it 
>> below the e.
>>
>> But 1.1.6 still clips the bulb.  It no longer transports the lost pixels 
>> to the other end of the string, but the clipping still happens.  Is 
>> there some new calling technique that I don't know about to get all of 
>> the text displayed?  I also see low descenders still being clipped (like 
>> the bottoms of g's).
>>     
>
> Sounds like you are running into the edge of the writing area,
> but expect PIL to ignore the clipping boundary for letters that
> are partially inside the clipped area.
>
> -- Scott David Daniels
> Scott.Daniels at Acm.Org
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>
>
>
>   

-- 
Ned Batchelder, http://nedbatchelder.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070110/16c157ef/attachment.htm 

From ned at nedbatchelder.com  Thu Jan 11 18:07:03 2007
From: ned at nedbatchelder.com (Ned Batchelder)
Date: Thu, 11 Jan 2007 12:07:03 -0500
Subject: [Image-SIG] PIL 1.1.6 eval function
In-Reply-To: <99ea8b7b0701110744l2e881938nd74fda2f27b93b5b@mail.gmail.com>
References: <4F19D8A3D7B26247A4481AE28348A848049FE0C3@gboexm2.ad.analog.com>	<99ea8b7b0701110743l744bfb4cyf086bf313a1b8cee@mail.gmail.com>
	<99ea8b7b0701110744l2e881938nd74fda2f27b93b5b@mail.gmail.com>
Message-ID: <45A66EB7.3030903@nedbatchelder.com>

The important thing to remember about eval() is that it does not operate 
on the entire pixel: the function is for a single color channel at a 
time, and behaves the same on all channels in the image.  So your 
function takes a single argument, and returns a single value.  Then the 
function is used to adjust the values of pixels one channel at a time.

Also, the function isn't actually evaluated for each pixel: it is used 
to build a lookup table, then the table is used to remap values.

--Ned.

Dennis Shimer wrote:
> This worked for me.
>
> open an image object, I'll call im
>
> Write a function that you would like applied to each pixel of im, here
> is a dumb one that will just double the value this making the new
> image very very bright.
>
> def function(pixel):
>     return pixel*2
>
> Then eval the image based on the function
>
> Image.eval(im2,function).show()
>
> or save it to a new image which you can save
>
> newimage=Image.eval(im2,function)
>
> Dennis Shimer
>
>
> On 1/11/07, Cochran, Travis <Travis.Cochran at analog.com> wrote:
>   
>> I'm trying to use the Image.eval function on a JPEG image. I can't seem
>> to find the right syntax for entering the function and it gives me
>> errors for the variables I try to use, "R,G,B". Any idea what I'm doing
>> wrong?
>> -Travis Cochran
>> _______________________________________________
>> Image-SIG maillist  -  Image-SIG at python.org
>> http://mail.python.org/mailman/listinfo/image-sig
>>
>>     
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>
>
>
>   

-- 
Ned Batchelder, http://nedbatchelder.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070111/968f6ff0/attachment.html 

From Scott.Daniels at Acm.Org  Tue Jan 16 05:06:25 2007
From: Scott.Daniels at Acm.Org (Scott David Daniels)
Date: Mon, 15 Jan 2007 20:06:25 -0800
Subject: [Image-SIG] Clipped text, still.
In-Reply-To: <45A52C83.2070102@nedbatchelder.com>
References: <45A40BE5.7010904@nedbatchelder.com> <eo1vsd$a2n$1@sea.gmane.org>
	<45A52C83.2070102@nedbatchelder.com>
Message-ID: <eohito$rm1$1@sea.gmane.org>

Ned Batchelder wrote:
>>> I tried installing 1.1.6, mostly to see the improvements in clipped 
>>> text.   PIL 1.1.5 would clip pieces of text ... and wrap those pixels....
>>> So "joe" might lose part of the bulb of the j and display it below the e.
>>>
>>> But 1.1.6 still clips the bulb.  It no longer transports the lost 
>>> pixels to the other end of the string....
 > Scott David Daniels wrote:
 >> Sounds like you are running into the edge of the writing area,
 >> but expect PIL to ignore the clipping boundary for letters that
 >> are partially inside the clipped area.

> In this case, there is no clipping boundary, and there are plenty of 
> pixels to accommodate the extremeties of the letters.  The wrapping has 
> been fixed, but the clipping still occurs.  Is this true or not?

Well, it sounds to me like you need to provide code, not simply
describe what might be going on.  There's no reason to debug a
description if you can make a simple small example (and that
example will be needed for a regression test in any case).

-- Scott David Daniels
Scott.Daniels at Acm.Org


From kemmerling at gmx.at  Tue Jan 16 06:57:32 2007
From: kemmerling at gmx.at (Markus Kemmerling)
Date: Tue, 16 Jan 2007 06:57:32 +0100
Subject: [Image-SIG] Tiff not reading dpi
In-Reply-To: <99ea8b7b0701110142g5e59ffa0t687eaaf71051b844@mail.gmail.com>
References: <99ea8b7b0701110142g5e59ffa0t687eaaf71051b844@mail.gmail.com>
Message-ID: <1B5C2AE3-FEAD-440F-8139-A10E62A76E70@gmx.at>

The Tiff image plugin handles the RESOLUTION_UNIT field incorrectly,  
see http://mail.python.org/pipermail/image-sig/2006-July/004011.html.

Markus Kemmerling
Medical University Vienna

Am 11.01.2007 um 10:42 schrieb Dennis Shimer:

> I have a tif image that at least 3 different programs can see the
> correct dpi information in
> (Irfanview, Gimp, Alchemy) but when I read the image using pil 1.6
>
> print im.info  -  shows only {'compression': 'raw'}
> dpix,dpiy=im.info['dpi']  -  gives an exception
>
> and if I set it explicitely then using
> im.save(filename,dpi=(dpix,dpiy))
>
> Irfanview sees the new dpi.
> Gimp complains that "resoultion units are meaningless"
> and Alchemy sees no DPI info at all but doesn't complain.
>
> Is there any reason that there is so much inconsistancy in results, or
> a way to write dpi with pil that would be universally recognized?
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig


From dshimer at gmail.com  Tue Jan 16 12:20:22 2007
From: dshimer at gmail.com (Dennis Shimer)
Date: Tue, 16 Jan 2007 06:20:22 -0500
Subject: [Image-SIG] Tiff not reading dpi
In-Reply-To: <1B5C2AE3-FEAD-440F-8139-A10E62A76E70@gmx.at>
References: <99ea8b7b0701110142g5e59ffa0t687eaaf71051b844@mail.gmail.com>
	<1B5C2AE3-FEAD-440F-8139-A10E62A76E70@gmx.at>
Message-ID: <99ea8b7b0701160320t60d8189bs89739d766238e96@mail.gmail.com>

Fantastic, thanks so much, that is exactly the fix I needed.

On 1/16/07, Markus Kemmerling <kemmerling at gmx.at> wrote:
> The Tiff image plugin handles the RESOLUTION_UNIT field incorrectly,
> see http://mail.python.org/pipermail/image-sig/2006-July/004011.html.
>
> Markus Kemmerling
> Medical University Vienna
>
> Am 11.01.2007 um 10:42 schrieb Dennis Shimer:
>
> > I have a tif image that at least 3 different programs can see the
> > correct dpi information in
> > (Irfanview, Gimp, Alchemy) but when I read the image using pil 1.6
> >
> > print im.info  -  shows only {'compression': 'raw'}
> > dpix,dpiy=im.info['dpi']  -  gives an exception
> >
> > and if I set it explicitely then using
> > im.save(filename,dpi=(dpix,dpiy))
> >
> > Irfanview sees the new dpi.
> > Gimp complains that "resoultion units are meaningless"
> > and Alchemy sees no DPI info at all but doesn't complain.
> >
> > Is there any reason that there is so much inconsistancy in results, or
> > a way to write dpi with pil that would be universally recognized?
> > _______________________________________________
> > Image-SIG maillist  -  Image-SIG at python.org
> > http://mail.python.org/mailman/listinfo/image-sig
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>

From ned at nedbatchelder.com  Tue Jan 16 13:10:02 2007
From: ned at nedbatchelder.com (Ned Batchelder)
Date: Tue, 16 Jan 2007 07:10:02 -0500
Subject: [Image-SIG] Clipped text, still.
In-Reply-To: <eohito$rm1$1@sea.gmane.org>
References: <45A40BE5.7010904@nedbatchelder.com>
	<eo1vsd$a2n$1@sea.gmane.org>	<45A52C83.2070102@nedbatchelder.com>
	<eohito$rm1$1@sea.gmane.org>
Message-ID: <45ACC09A.9040404@nedbatchelder.com>

OK, here's some code, and some output.  (My emails seem to be 
consistently held for moderation; I apologize for the apparent delays in 
responding).

This code (in a Django view function):

    im = Image.new("RGB", (300, 300), '#ffffc0')
    draw = ImageDraw.Draw(im)
    font = ImageFont.truetype('timesbi.ttf', 50)
    draw.text((100, 100), "jigf", font=font, fill='black')
    response = HttpResponse(mimetype='image/png')
    im.save(response, 'PNG')
    return response

produces two different results, attached.  One is with PIL 1.1.5, the 
other with 1.1.6.  1.1.5 produces the text in the proper place, but with 
the bulbs of the j and f wrapped around to the other end of the text.  
1.1.6 doesn't wrap the bulbs, but has clipped the j severely, and has 
mis-positioned the text so as to keep the entire f in the drawing.

Is this not something other people are seeing?

--Ned.

Scott David Daniels wrote:
> Ned Batchelder wrote:
>   
>>>> I tried installing 1.1.6, mostly to see the improvements in clipped 
>>>> text.   PIL 1.1.5 would clip pieces of text ... and wrap those pixels....
>>>> So "joe" might lose part of the bulb of the j and display it below the e.
>>>>
>>>> But 1.1.6 still clips the bulb.  It no longer transports the lost 
>>>> pixels to the other end of the string....
>>>>         
>  > Scott David Daniels wrote:
>  >> Sounds like you are running into the edge of the writing area,
>  >> but expect PIL to ignore the clipping boundary for letters that
>  >> are partially inside the clipped area.
>
>   
>> In this case, there is no clipping boundary, and there are plenty of 
>> pixels to accommodate the extremeties of the letters.  The wrapping has 
>> been fixed, but the clipping still occurs.  Is this true or not?
>>     
>
> Well, it sounds to me like you need to provide code, not simply
> describe what might be going on.  There's no reason to debug a
> description if you can make a simple small example (and that
> example will be needed for a regression test in any case).
>
> -- Scott David Daniels
> Scott.Daniels at Acm.Org
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>
>
>
>   

-- 
Ned Batchelder, http://nedbatchelder.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070116/ed6c3a28/attachment-0001.htm 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: jigf_116.png
Type: image/png
Size: 6947 bytes
Desc: not available
Url : http://mail.python.org/pipermail/image-sig/attachments/20070116/ed6c3a28/attachment-0002.png 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: jigf_115.png
Type: image/png
Size: 7270 bytes
Desc: not available
Url : http://mail.python.org/pipermail/image-sig/attachments/20070116/ed6c3a28/attachment-0003.png 

From Scott.Daniels at Acm.Org  Tue Jan 16 16:47:57 2007
From: Scott.Daniels at Acm.Org (Scott David Daniels)
Date: Tue, 16 Jan 2007 07:47:57 -0800
Subject: [Image-SIG] Interesting Bounding box calculation
In-Reply-To: <45ACC09A.9040404@nedbatchelder.com>
References: <45A40BE5.7010904@nedbatchelder.com>	<eo1vsd$a2n$1@sea.gmane.org>	<45A52C83.2070102@nedbatchelder.com>	<eohito$rm1$1@sea.gmane.org>
	<45ACC09A.9040404@nedbatchelder.com>
Message-ID: <eois14$a6p$1@sea.gmane.org>

Ned Batchelder wrote:
> OK, here's some code, and some output....

After looking, and doing a bit of experimenting, here is a
surprising drawing:

     import Image, ImageDraw, ImageFont
     def paints(fontname='timesbi.ttf', fontsize=50,
               canvas=(500, 700), position=(100, 100)):
         im = Image.new("RGB", canvas, '#ffffc0')
         draw = ImageDraw.Draw(im)
         font = ImageFont.truetype(fontname, fontsize)
         fh, fx = font.getmetrics()
         x, y = position
         for text in ['j_0_|igf', ' j_1_|igf', '  j_2_|igf', 'young']:
	    width, height = font.getsize(text) # w,h
             draw.rectangle((x, y, x+width, y+height),
                            fill='lightgrey', outline='lightgreen')
             draw.rectangle((x-fx, y, x, y+fh),
                            fill='lightgrey', outline='red')
             draw.text((x, y), text, font=font, fill='black')
             y += 100
         return im
     paints().show()

-- Scott David Daniels
Scott.Daniels at Acm.Org


From rowen at cesmail.net  Tue Jan 16 22:32:55 2007
From: rowen at cesmail.net (Russell E Owen)
Date: Tue, 16 Jan 2007 13:32:55 -0800
Subject: [Image-SIG] PIL 1.1.6 gotcha; Image.frombuffer
Message-ID: <rowen-01EF90.13325516012007@sea.gmane.org>

My python code started generating warnings when used with PIL 1.1.6 in 
Image.frombuffer. The warning says to specify all arguments, to avoid 
problems when a default will change. But either the warning is wrong or 
I misread it because I thought it said:

Image.frombuffer(mode, shape, data, "raw", mode, 0, 1)

But when I did this my images were upside down. The final 1 should be -1 
if you want the old default behavior. In other words:

Image.frombuffer(mode, size, data)

should be changed to:

Image.frombuffer(mode, size, data, "raw", mode, 0, -1)

-- Russell

P.S. it would be a big help if the manual were explicit about what the 
old default orientation was for frombuffer (beyond just warning that it 
may change).


From ned at nedbatchelder.com  Tue Jan 16 19:27:05 2007
From: ned at nedbatchelder.com (Ned Batchelder)
Date: Tue, 16 Jan 2007 13:27:05 -0500
Subject: [Image-SIG] Interesting Bounding box calculation
In-Reply-To: <eois14$a6p$1@sea.gmane.org>
References: <45A40BE5.7010904@nedbatchelder.com>	<eo1vsd$a2n$1@sea.gmane.org>	<45A52C83.2070102@nedbatchelder.com>	<eohito$rm1$1@sea.gmane.org>	<45ACC09A.9040404@nedbatchelder.com>
	<eois14$a6p$1@sea.gmane.org>
Message-ID: <45AD18F9.50406@nedbatchelder.com>

I'm not sure exactly what you are showing me here.  Perhaps you could 
describe what you see (or even post the images).  Prefixing the string 
with spaces will of course prevent the problem, but introduces problems 
of its own (how to accurately position the text, how to avoid the spaces 
clobbering other text, etc).  You're using font.getmetrics() here, which 
is an undocumented function.  Reading the code, it seems to return the 
ascent and descent of the font, so using either dimension for an x value 
seems wrong to me, and I'm not sure why you are using it in the first place.

I want to understand.  Can you explain?

--Ned.

Scott David Daniels wrote:
> Ned Batchelder wrote:
>   
>> OK, here's some code, and some output....
>>     
>
> After looking, and doing a bit of experimenting, here is a
> surprising drawing:
>
>      import Image, ImageDraw, ImageFont
>      def paints(fontname='timesbi.ttf', fontsize=50,
>                canvas=(500, 700), position=(100, 100)):
>          im = Image.new("RGB", canvas, '#ffffc0')
>          draw = ImageDraw.Draw(im)
>          font = ImageFont.truetype(fontname, fontsize)
>          fh, fx = font.getmetrics()
>          x, y = position
>          for text in ['j_0_|igf', ' j_1_|igf', '  j_2_|igf', 'young']:
> 	    width, height = font.getsize(text) # w,h
>              draw.rectangle((x, y, x+width, y+height),
>                             fill='lightgrey', outline='lightgreen')
>              draw.rectangle((x-fx, y, x, y+fh),
>                             fill='lightgrey', outline='red')
>              draw.text((x, y), text, font=font, fill='black')
>              y += 100
>          return im
>      paints().show()
>
> -- Scott David Daniels
> Scott.Daniels at Acm.Org
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>
>
>
>   

-- 
Ned Batchelder, http://nedbatchelder.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070116/a7206c7e/attachment.htm 

From wrybread at gmail.com  Thu Jan 18 01:50:06 2007
From: wrybread at gmail.com (Alec Bennett)
Date: Wed, 17 Jan 2007 16:50:06 -0800
Subject: [Image-SIG] rotate image with transparent background?
Message-ID: <cab7082b0701171650h5238786o91027a55d3e7eb1d@mail.gmail.com>

I'm trying to rotate an image and have the expanded background be transparent.

For example:

im = Image.open("pic1.jpg")
im2 = im.rotate(32, expand = 1)
im2.show()

In the above, im2 is a copy of pic1.jpg that has been rotated 32
degrees. The problem is that the background over where it was rotated
is a black box.

Is there a way to make that background transparent? I'll be pasting
the image onto another image, and don't want a bunch of black shapes.

Thanks for any help.

From douglas at paradise.net.nz  Thu Jan 18 02:55:08 2007
From: douglas at paradise.net.nz (Douglas Bagnall)
Date: Thu, 18 Jan 2007 14:55:08 +1300
Subject: [Image-SIG] rotate image with transparent background?
In-Reply-To: <cab7082b0701171650h5238786o91027a55d3e7eb1d@mail.gmail.com>
References: <cab7082b0701171650h5238786o91027a55d3e7eb1d@mail.gmail.com>
Message-ID: <45AED37C.5020306@paradise.net.nz>

hi Alec,

> I'm trying to rotate an image and have the expanded background be transparent.

The image you get from opening a JPG won't have an alpha channel. You
need to add it yourself with the .convert() method.

This ought to work:

im = Image.open("pic1.jpg").convert('RGBA')
im2 = im.rotate(32, expand = 1)
im2.show()



Douglas Bagnall

From wrybread at gmail.com  Thu Jan 18 04:55:33 2007
From: wrybread at gmail.com (Alec Bennett)
Date: Wed, 17 Jan 2007 19:55:33 -0800
Subject: [Image-SIG] rotate image with transparent background?
In-Reply-To: <45AED37C.5020306@paradise.net.nz>
References: <cab7082b0701171650h5238786o91027a55d3e7eb1d@mail.gmail.com>
	<45AED37C.5020306@paradise.net.nz>
Message-ID: <cab7082b0701171955j6e5e5df9vcbb9757dda3fe99c@mail.gmail.com>

Hmm, still getting a black box. The code as it stands now in case
anything jumps out at you:


from PIL import Image

collage = Image.open("collage.jpg")

# open the pic and give it an alpha channel so it's transparent
im1 = Image.open("pic1.jpg")
im1.convert('RGBA')

# rotate it and expand it's canvass so the corners don't get cut off:
im2 = im1.rotate(32, expand = 1)

collage.paste(im2, (30, 30) )
collage.save("collage.jpg")

collage.show()




On 1/17/07, Douglas Bagnall <douglas at paradise.net.nz> wrote:
> hi Alec,
>
> > I'm trying to rotate an image and have the expanded background be transparent.
>
> The image you get from opening a JPG won't have an alpha channel. You
> need to add it yourself with the .convert() method.
>
> This ought to work:
>
> im = Image.open("pic1.jpg").convert('RGBA')
> im2 = im.rotate(32, expand = 1)
> im2.show()
>
>
>
> Douglas Bagnall
>

From douglas at paradise.net.nz  Thu Jan 18 05:13:15 2007
From: douglas at paradise.net.nz (Douglas Bagnall)
Date: Thu, 18 Jan 2007 17:13:15 +1300
Subject: [Image-SIG] rotate image with transparent background?
In-Reply-To: <cab7082b0701171955j6e5e5df9vcbb9757dda3fe99c@mail.gmail.com>
References: <cab7082b0701171650h5238786o91027a55d3e7eb1d@mail.gmail.com>
	<45AED37C.5020306@paradise.net.nz>
	<cab7082b0701171955j6e5e5df9vcbb9757dda3fe99c@mail.gmail.com>
Message-ID: <45AEF3DB.207@paradise.net.nz>

Alec Bennett wrote:
> Hmm, still getting a black box. The code as it stands now in case
> anything jumps out at you:

> # open the pic and give it an alpha channel so it's transparent
> im1 = Image.open("pic1.jpg")
> im1.convert('RGBA')

That should be

im1 = Image.open("pic1.jpg")
im1 = im1.convert('RGBA')
^^^^^
or more concisely

im1 = Image.open("pic1.jpg").convert('RGBA')

Like most PIL methods, convert() returns the modified image, leaving the
original one as is.


douglas


From douglas at paradise.net.nz  Thu Jan 18 05:56:06 2007
From: douglas at paradise.net.nz (Douglas Bagnall)
Date: Thu, 18 Jan 2007 17:56:06 +1300
Subject: [Image-SIG] rotate image with transparent background?
In-Reply-To: <cab7082b0701172032of204166ucf48db45757708da@mail.gmail.com>
References: <cab7082b0701171650h5238786o91027a55d3e7eb1d@mail.gmail.com>
	<45AED37C.5020306@paradise.net.nz>
	<cab7082b0701171955j6e5e5df9vcbb9757dda3fe99c@mail.gmail.com>
	<45AEF3DB.207@paradise.net.nz>
	<cab7082b0701172032of204166ucf48db45757708da@mail.gmail.com>
Message-ID: <45AEFDE6.7030701@paradise.net.nz>

Alec Bennett wrote:

> collage.paste(im2, (30, 30) )

>From http://effbot.org/imagingbook/image.htm#Image.paste :

  im.paste(image, box, mask)

  Same as above, but updates only the regions indicated by the
  mask. You can use either "1", "L" or "RGBA" images (in the latter
  case, the alpha band is used as mask). Where the mask is 255, the
  given image is copied as is. Where the mask is 0, the current value
  is preserved. Intermediate values can be used for transparency
  effects.

  Note that if you paste an "RGBA" image, the alpha band is
  ignored. You can work around this by using the same image as both
  source image and mask.

so you want

collage.paste(im2, (30, 30), im2)



douglas

From wrybread at gmail.com  Thu Jan 18 06:19:21 2007
From: wrybread at gmail.com (Alec Bennett)
Date: Wed, 17 Jan 2007 21:19:21 -0800
Subject: [Image-SIG] rotate image with transparent background?
In-Reply-To: <45AED37C.5020306@paradise.net.nz>
References: <cab7082b0701171650h5238786o91027a55d3e7eb1d@mail.gmail.com>
	<45AED37C.5020306@paradise.net.nz>
Message-ID: <cab7082b0701172119u6b32091el440bd02b73dda8ee@mail.gmail.com>

Douglas helped me off list, wanted to post the complete code in case
anyone else comes looking.  So here's the code to rotate an image and
paste it onto another image without creating a black bounding box:


from PIL import Image

collage = Image.open("cheese.jpg")

# open the pic and give it an alpha channel so it's transparent
im1 = Image.open("pic1.jpg").convert('RGBA')

# rotate it and expand it's canvas so the corners don't get cut off:
im2 = im1.rotate(32, expand = 1)

# note the second appearance of im2, that's necessary to paste without a bg
collage.paste(im2, (30, 30), im2 )
collage.save("new.jpg")

collage.show()

From mypassion_13 at yahoo.co.in  Thu Jan 18 15:18:31 2007
From: mypassion_13 at yahoo.co.in (sanket jalan)
Date: Thu, 18 Jan 2007 14:18:31 +0000 (GMT)
Subject: [Image-SIG]  Image matching
Message-ID: <20070118141831.50479.qmail@web8915.mail.in.yahoo.com>

Hi there !!

Actually I am new to the field of Digital Image processing and I have to make a project in which I have to compare any input images with those in the database and give results according to the degree of similarity. The images can be any geographical maps, faces etc. but a grayscale only at this stage. 
Can you help me to know what series of techniques I must use for the purpose. As I was thinking to apply edge detection then thresholding and after that comparison of this image. but I am not sure. Is this the best possible way? 


Thanx in advace
Warm Regards
Sanket Jalan 
 				
---------------------------------
 Here?s a new way to find what you're looking for - Yahoo! Answers 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070118/f055b7b2/attachment.htm 

From rsenra at acm.org  Sat Jan 20 17:12:55 2007
From: rsenra at acm.org (Rodrigo Senra)
Date: Sat, 20 Jan 2007 14:12:55 -0200
Subject: [Image-SIG] Image matching
In-Reply-To: <20070118141831.50479.qmail@web8915.mail.in.yahoo.com>
References: <20070118141831.50479.qmail@web8915.mail.in.yahoo.com>
Message-ID: <6C496C4E-DA6E-4D93-9F5B-BD197B049FB7@acm.org>


On 18Jan 2007, at 12:18 PM, sanket jalan wrote:

> Hi there !!
>
> Actually I am new to the field of Digital Image processing and I  
> have to make a project in which I have to compare any input images  
> with those in the database and give results according to the degree  
> of similarity.

Color similarity, shape similarity, semantics ?

Try studying these python-based projects:

- ImgSeek [1]
   """imgSeek is a photo collection manager and viewer with
   content-based search and many other features."""

- SOM [2, 3]

For a broader view, check the Wikipedia [4].

[1] http://www.imgseek.net/
[2] http://rtshiva.wordpress.com/2006/08/27/som-content-based-image- 
retrieval/
[3] http://generation5.org/content/2004/aiSomPic.asp
[4] http://en.wikipedia.org/wiki/CBIR

>  As I was thinking to apply edge detection then thresholding and  
> after that comparison of this image. but I am not sure. Is this the  
> best possible way?

Even though DIP is not my field, I beleive content-based image  
retrieval is
still an open problem.

HTH,
Senra


Rodrigo Senra
______________
rsenra @ acm.org
http://rodrigo.senra.nom.br





From Travis.Cochran at analog.com  Mon Jan 22 14:40:44 2007
From: Travis.Cochran at analog.com (Cochran, Travis)
Date: Mon, 22 Jan 2007 08:40:44 -0500
Subject: [Image-SIG] Creating a grayscale ramp
Message-ID: <4F19D8A3D7B26247A4481AE28348A84804A87DDB@gboexm2.ad.analog.com>

I'm trying to use the PIL to create a grayscale ramp, dark to light. I
am able to generate an image but it is solid black. Am I using
frombuffer wrong? Here is my code.

from PIL import Image

#Creates the list of codes from 0 to 255
data = range(256) 
#Creates an image out of those codes
im = Image.frombuffer("L", (256, 1), str(data), "raw", "L", 0, 1)
#Resizes the image to make it easy to view
im = im.resize((256, 768)).show()

From export at hope.cz  Mon Jan 22 15:11:38 2007
From: export at hope.cz (export at hope.cz)
Date: Mon, 22 Jan 2007 15:11:38 +0100
Subject: [Image-SIG] How to write a text to an image
Message-ID: <45B4D42A.25963.165AF45A@export.hope.cz>

Is it possible to write a text , by using PIL, to a bottom corner   of an image?
Is there any doc/tutorial  for that?
Thank you
Lad.


From rsenra at acm.org  Mon Jan 22 16:01:42 2007
From: rsenra at acm.org (Rodrigo Senra)
Date: Mon, 22 Jan 2007 13:01:42 -0200
Subject: [Image-SIG] How to write a text to an image
In-Reply-To: <45B4D42A.25963.165AF45A@export.hope.cz>
References: <45B4D42A.25963.165AF45A@export.hope.cz>
Message-ID: <709BBB76-CBC7-4A6A-81C0-5F93F90C896D@acm.org>


On 22Jan 2007, at 12:11 PM, export at hope.cz wrote:

> Is it possible to write a text , by using PIL, to a bottom corner
>   of an image?
> Is there any doc/tutorial  for that?

I have created this recipe [1] some time ago, I guess this it is what  
(or
something very close to what) you are looking for.
There are other useful recipes there.

[1] http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/483756

best regards,
Senra


Rodrigo Senra
______________
rsenra @ acm.org
http://rodrigo.senra.nom.br





From douglas at paradise.net.nz  Mon Jan 22 20:57:10 2007
From: douglas at paradise.net.nz (Douglas Bagnall)
Date: Tue, 23 Jan 2007 08:57:10 +1300
Subject: [Image-SIG] Creating a grayscale ramp
In-Reply-To: <4F19D8A3D7B26247A4481AE28348A84804A87DDB@gboexm2.ad.analog.com>
References: <4F19D8A3D7B26247A4481AE28348A84804A87DDB@gboexm2.ad.analog.com>
Message-ID: <45B51716.3080407@paradise.net.nz>

Cochran, Travis wrote:
> I'm trying to use the PIL to create a grayscale ramp, dark to light. I
> am able to generate an image but it is solid black. Am I using
> frombuffer wrong? Here is my code.
> 
> from PIL import Image
> 
> #Creates the list of codes from 0 to 255
> data = range(256) 
> #Creates an image out of those codes
> im = Image.frombuffer("L", (256, 1), str(data), "raw", "L", 0, 1)
                                       ^^^^^^^^^
str(data) turns the list into a 1170 byte string representation of the
list, beginning "[0, 1, 2,".  Translated back to byte values, this is
[91, 48, 44, 32, 49, 44, 32, 50, 44, 32], not [0, 1, 2].  Your "solid
black" is actually a dappled dark grey.

What you want is a 256 byte string of ascending values, which would
begin (in python's string notation) "\x00\x01\x02".  One way to get this
would be:

# you can leave out the square brackets for python 2.4 +
data = ''.join([chr(x) for x in range(256)])

im = Image.frombuffer("L", (256, 1), data, "raw", "L", 0, 1)


Note, here:
> #Resizes the image to make it easy to view
> im = im.resize((256, 768)).show()

you are setting im to None, because that is what .show() returns.  You
probably don't mean this, and would be better using two lines:

im = im.resize((256, 768))
im.show()




douglas



From carroll at tjc.com  Wed Jan 24 07:10:32 2007
From: carroll at tjc.com (Terry Carroll)
Date: Tue, 23 Jan 2007 22:10:32 -0800 (PST)
Subject: [Image-SIG] How to write a text to an image
In-Reply-To: <45B4D42A.25963.165AF45A@export.hope.cz>
Message-ID: <Pine.LNX.4.44.0701232204470.17946-100000@violet.rahul.net>

On Mon, 22 Jan 2007 export at hope.cz wrote:

> Is it possible to write a text , by using PIL, to a bottom corner of an
> image? Is there any doc/tutorial for that?

I use this routine (in a little app I wrote to extract the date a photo
was taken from its EXIF data, and add it to the bottom corner)

def Imprint(im, inputtext, font=None, color=None, opacity=.6, margin=(30,30)):
    """
    imprints a PIL image with the indicated text in lower-right corner
    """
    if im.mode != "RGBA":
        im = im.convert("RGBA")
    textlayer = Image.new("RGBA", im.size, (0,0,0,0))
    textdraw = ImageDraw.Draw(textlayer)
    textsize = textdraw.textsize(inputtext, font=font)
    textpos = [im.size[i]-textsize[i]-margin[i] for i in [0,1]]
    textdraw.text(textpos, inputtext, font=font, fill=color)
    if opacity != 1:
        textlayer = ReduceOpacity(textlayer,opacity)
    return Image.composite(textlayer, im, textlayer)

The ReduceOpacity method is the one from this recipe:

http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/362879

Which inspired my app in the first place


From jwhiteillustration at cox.net  Wed Jan 24 16:37:31 2007
From: jwhiteillustration at cox.net (Janie White)
Date: Wed, 24 Jan 2007 08:37:31 -0700
Subject: [Image-SIG] leptoprin scam zantrex scam strivectin scam anorex scam
Message-ID: <5246DCEF-DAD2-4BCA-A6B0-E84A23BD7CE2@cox.net>

Hello,
Who are you and how do I know that your information is accurate?
My mother is taken in by the "Better than Botox?" ad- I thought that  
I'd do some research- What you have posted sounds on target, but how  
do I verify your information?
Thanks,
Janie


From ken at seehart.com  Wed Jan 24 20:51:46 2007
From: ken at seehart.com (Ken Seehart)
Date: Wed, 24 Jan 2007 11:51:46 -0800
Subject: [Image-SIG] video capture
Message-ID: <45B7B8D2.4020500@seehart.com>

Anyone know what is the closest to the following?  I am looking for 
video capture functionality.

I am developing a  multi platform application written in wxPython.  The 
feature I need to add is video conferencing and streaming.  
Specifically, I need to create a window class that can receive video 
from our server or from a peer.  I also need video capture and the 
ability to output a video stream to one or more peers.

I think maybe Kamelia could do the pipelining, but they don't seem to 
have much video capture capability built in.  Unless Twisted is 
better... opinions?
In any case, I need video capture, which I probably can wrap into Kamelia.

The only interesting hits I have found so far are:
PXC-200 Frame Grabber Module - http://www.mems-exchange.org/software/pxc200/
libfg -- Frame Grabber library for Linux - http://antonym.org/libfg

These are all either camera hardware specific or platform specific or 
both.  Isn't there an open source multi-platform library for this kind 
of thing?

If such a thing exists in C++, perhaps it could be swigificated.

BTW, there is probably a short term contract available for anyone with 
the ability to develop this functionality, and I might be able to 
arrange for the resulting module to be open source (subject to 
approval).  Contact me off line if you are interested in that.

- Ken










From dshimer at gmail.com  Wed Jan 24 21:10:03 2007
From: dshimer at gmail.com (Dennis Shimer)
Date: Wed, 24 Jan 2007 15:10:03 -0500
Subject: [Image-SIG] video capture
In-Reply-To: <45B7B8D2.4020500@seehart.com>
References: <45B7B8D2.4020500@seehart.com>
Message-ID: <99ea8b7b0701241210k29fd390fy10ff7f3907b84dc2@mail.gmail.com>

Have you looked at what makes virtualdub tick?  I have used it as a
capture app for quite some time and am amazed at the function and
efficiency of it.  If memory serves it is C++ and open source.

From ken at seehart.com  Wed Jan 24 21:44:53 2007
From: ken at seehart.com (Ken Seehart)
Date: Wed, 24 Jan 2007 12:44:53 -0800
Subject: [Image-SIG] video capture
In-Reply-To: <99ea8b7b0701241210k29fd390fy10ff7f3907b84dc2@mail.gmail.com>
References: <45B7B8D2.4020500@seehart.com>
	<99ea8b7b0701241210k29fd390fy10ff7f3907b84dc2@mail.gmail.com>
Message-ID: <45B7C545.1010609@seehart.com>

Dennis Shimer wrote:
> Have you looked at what makes virtualdub tick?  I have used it as a
> capture app for quite some time and am amazed at the function and
> efficiency of it.  If memory serves it is C++ and open source.
>
>
Thanks, that looks nice.  I might dig into it more if I can't find a 
better match, but virtualdib is Windows only, and not packaged as a library.

- Ken



From seb.haase at gmx.net  Wed Jan 24 22:48:03 2007
From: seb.haase at gmx.net (Sebastian Haase)
Date: Wed, 24 Jan 2007 13:48:03 -0800
Subject: [Image-SIG] video capture
In-Reply-To: <45B7C545.1010609@seehart.com>
References: <45B7B8D2.4020500@seehart.com>
	<99ea8b7b0701241210k29fd390fy10ff7f3907b84dc2@mail.gmail.com>
	<45B7C545.1010609@seehart.com>
Message-ID: <bc657ead0701241348r175f006ase7e6235cec358efd@mail.gmail.com>

Do you know about
VideoCapture
  A Win32 Python Extension for Accessing Video Devices (e.g. a USB
WebCam, a TV-Card, ...)
   by Markus Gritsch (gritsch at iue.tuwien.ac.at)

-Sebastian Haase


On 1/24/07, Ken Seehart <ken at seehart.com> wrote:
> Dennis Shimer wrote:
> > Have you looked at what makes virtualdub tick?  I have used it as a
> > capture app for quite some time and am amazed at the function and
> > efficiency of it.  If memory serves it is C++ and open source.
> >
> >
> Thanks, that looks nice.  I might dig into it more if I can't find a
> better match, but virtualdib is Windows only, and not packaged as a library.
>
> - Ken
>
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>

From nicolas at famillepinault.fr  Thu Jan 25 07:43:50 2007
From: nicolas at famillepinault.fr (Nicolas Pinault)
Date: Thu, 25 Jan 2007 07:43:50 +0100
Subject: [Image-SIG] video capture
In-Reply-To: <bc657ead0701241348r175f006ase7e6235cec358efd@mail.gmail.com>
References: <45B7B8D2.4020500@seehart.com>	<99ea8b7b0701241210k29fd390fy10ff7f3907b84dc2@mail.gmail.com>	<45B7C545.1010609@seehart.com>
	<bc657ead0701241348r175f006ase7e6235cec358efd@mail.gmail.com>
Message-ID: <45B851A6.8050706@famillepinault.fr>



Sebastian Haase a ?crit :
> Do you know about
> VideoCapture
>   A Win32 Python Extension for Accessing Video Devices (e.g. a USB
> WebCam, a TV-Card, ...)
>    by Markus Gritsch (gritsch at iue.tuwien.ac.at)
>
> -Sebastian Haase
>   
Video capture does not capture streams. Videocapture only grabs single 
frames. It is usefull for webcam for example where the app grabs a 
picture every minute or every second or whatever you want. You can also 
grabs picture at a higher rate than the video stream, doubling or 
tripling source frames.

Nicolas

From p.lists at partylemon.com  Thu Jan 25 08:27:02 2007
From: p.lists at partylemon.com (Philip Quinn)
Date: Thu, 25 Jan 2007 20:27:02 +1300
Subject: [Image-SIG] PIL stripping colour profile
Message-ID: <EBDA8E8F-52C2-4B80-9F5B-8F0673E21C1D@partylemon.com>

Hi,

I've noticed that if I load an image with an embedded colour profile  
with PIL and then save it out again, then colour profile is lost.
Is it possible to either stop PIL from doing this, or re-attach the  
profile to the image?

Thanks,

Phil


From PMF at kingswood.bath.sch.uk  Thu Jan 25 11:52:18 2007
From: PMF at kingswood.bath.sch.uk (Paul)
Date: Thu, 25 Jan 2007 10:52:18 -0000
Subject: [Image-SIG] Newb help... can't install the PIL extension libraries
Message-ID: <06C7A71EB82A7E4F8FBB65E016C15D07072A3B@kws-exc-01.kingswood.internal>

Hi,

 

I'm very new to linux (2 days!) but have been using Python on windows
for a few weeks now. I would like to install PIL on my linux (ubuntu)
machine but am having all sorts of trouble following the readme, I get a
ton of errors when trying to run the build.py. Mostly cannot find
directory errors, but there are others thrown in for good measure.

 

This is my first attempt at importing a library within linux.

 

I unzipped the tar file into my user space...

 

/home/paul/Python/ExtensionLibraries

 

which I created myself. This is where I'm going wrong I presume. The
readme said that I could create my own directory, so I'm confused.

 

Should I unzip it & build it somewhere else?

 

Any help greatly appreciated, many thanks.

 

PMF

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070125/f73d9994/attachment.html 

From Chris.Barker at noaa.gov  Thu Jan 25 20:41:38 2007
From: Chris.Barker at noaa.gov (Christopher Barker)
Date: Thu, 25 Jan 2007 11:41:38 -0800
Subject: [Image-SIG] Newb help... can't install the PIL extension
	libraries
In-Reply-To: <06C7A71EB82A7E4F8FBB65E016C15D07072A3B@kws-exc-01.kingswood.internal>
References: <06C7A71EB82A7E4F8FBB65E016C15D07072A3B@kws-exc-01.kingswood.internal>
Message-ID: <45B907F2.7010104@noaa.gov>

Paul wrote:
> I?m very new to linux (2 days!) but have been using Python on windows 
> for a few weeks now. I would like to install PIL on my linux (ubuntu) 
> machine but am having all sorts of trouble following the readme, I get a 
> ton of errors when trying to run the build.py. Mostly cannot find 
> directory errors, but there are others thrown in for good measure.

post a few here -- particularly the first few you get.

> I unzipped the tar file into my user space?
> /home/paul/Python/ExtensionLibraries

> which I created myself. This is where I?m going wrong I presume. The 
> readme said that I could create my own directory, so I?m confused.

> Should I unzip it & build it somewhere else?

no, that's not it, you can built it anywhere.

I'm guessing that you're missing some libraries -- you need the ones 
that PIL needs (libjpeg, freetype2, zlib). These are probably already on 
your system, but you may not have the "devel" versions, which you need 
as PIL is building against them. Sorry I'm not familiar with Ubuntu, but 
look for things like:

python-devel
jpeg-devel
zlib-devel

etc...

oh, and are you sure there is not an existing Ubuntu package for PIL?

-Chris


-- 
Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R            (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chris.Barker at noaa.gov


From ced at bernispa.com  Fri Jan 26 13:41:47 2007
From: ced at bernispa.com (Cesare Leonardi)
Date: Fri, 26 Jan 2007 13:41:47 +0100
Subject: [Image-SIG] Problem with JPEG and CMYK color space
In-Reply-To: <443A217C.1050405@bernispa.com>
References: <443A217C.1050405@bernispa.com>
Message-ID: <45B9F70B.1050806@bernispa.com>

Berni CED ha scritto:
> So i think there is a bug in the code related to JPEG CYMK or there are
> some JPEG unsupported feature. What do you think? Some advice?

Now PIL 1.1.6 includes a solution to that problem.

Since some days ago i was using PIL 1.1.4 with the patch from Kevin 
Cazabon and for my experiments everything was fine. Some days ago i've 
tryed to install python 2.5 along with PIL 1.1.6 (under Windows) and so 
i've tryed to test some of the JPEG CMYK files that was problematic with 
official PIL 1.1.4 and i noted that some images now are threated good 
but other still looks bad. Substituting "JpegImagePlugin.py" with the 
one patched by Kevin, everything is fine.
So seems that the file that comes with 1.1.6 is not functional 
equivalent to the one patched by Kevin.

I've posted some examples in the following site (see PART 5 at the end):
http://www.bernispa.com/pil/index.html

I've also attached to this message the diff between the 
"JpegImagePlugin.py" as found in PIL 1.1.6 and the same file patched
by Kevin (it was written for 1.1.4 but seems that the original file
as not changed meanwhile)

What do you think?

Cesare.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: diff.txt
Url: http://mail.python.org/pipermail/image-sig/attachments/20070126/3c75ea3b/attachment.txt 

From ken at seehart.com  Thu Jan 25 22:55:10 2007
From: ken at seehart.com (Ken Seehart)
Date: Thu, 25 Jan 2007 13:55:10 -0800
Subject: [Image-SIG] video capture
In-Reply-To: <45B851A6.8050706@famillepinault.fr>
References: <45B7B8D2.4020500@seehart.com>	<99ea8b7b0701241210k29fd390fy10ff7f3907b84dc2@mail.gmail.com>	<45B7C545.1010609@seehart.com>	<bc657ead0701241348r175f006ase7e6235cec358efd@mail.gmail.com>
	<45B851A6.8050706@famillepinault.fr>
Message-ID: <45B9273E.8010006@seehart.com>

Nicolas Pinault wrote:
> Sebastian Haase a ?crit :
>   
>> Do you know about
>> VideoCapture
>>   A Win32 Python Extension for Accessing Video Devices (e.g. a USB
>> WebCam, a TV-Card, ...)
>>    by Markus Gritsch (gritsch at iue.tuwien.ac.at)
>>
>> -Sebastian Haase
>>   
>>     
> Video capture does not capture streams. Videocapture only grabs single 
> frames. It is usefull for webcam for example where the app grabs a 
> picture every minute or every second or whatever you want. You can also 
> grabs picture at a higher rate than the video stream, doubling or 
> tripling source frames.
>
> Nicolas
>   
I'm not sure I understand the distinction.  For example, if I pipeline a 
series of camera images
at the desired frame rate into a component that converts a sequences of 
frames into a video
stream, then I achieve the desired result right?

I imagine that the video streaming component would surely be very 
portable, but the video
capture would tend to be O/S and hardware specific. 

Maybe what the world needs is a set of webcam drivers wrapped up into a 
single platform-
independent PIL compatible python module, that could then be wrapped 
into a Kamelia
component.  Anyone else interested in such a project?  I can't believe 
such a thing doesn't
already exist.

- Ken

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070125/dab11b54/attachment.htm 

From jason at smileproject.com  Thu Jan 25 23:17:48 2007
From: jason at smileproject.com (Jason Van Anden)
Date: Thu, 25 Jan 2007 17:17:48 -0500
Subject: [Image-SIG] Splitting Color Stacked TIFFs Problem
Message-ID: <8eadd6950701251417r44baf177u54a35ee3aa110810@mail.gmail.com>

Hello List,

A project I am developing in PIL splits stacked Tiffs (among other things).
It works super on B&W Tiffs, but chokes on color.

The error is "cannot identify image file".

My question is ... is this a limitation of PIL (Color Tiffs) or could
something be screwy about my files?  I am using the same program to capture
the images - just two different cameras.

Thank You,
Jason Van Anden
http://www.smileproject.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070125/4cf5dc91/attachment.html 

From waldemar.osuch at gmail.com  Fri Jan 26 16:57:01 2007
From: waldemar.osuch at gmail.com (Waldemar Osuch)
Date: Fri, 26 Jan 2007 08:57:01 -0700
Subject: [Image-SIG] Newb help... can't install the PIL extension
	libraries
In-Reply-To: <06C7A71EB82A7E4F8FBB65E016C15D07072A3B@kws-exc-01.kingswood.internal>
References: <06C7A71EB82A7E4F8FBB65E016C15D07072A3B@kws-exc-01.kingswood.internal>
Message-ID: <6fae95540701260757hd95bb58rfc62e5a04857c8f3@mail.gmail.com>

On 1/25/07, Paul <PMF at kingswood.bath.sch.uk> wrote:
>
>
>
>
> Hi,
>
>
>
> I'm very new to linux (2 days!) but have been using Python on windows for a
> few weeks now. I would like to install PIL on my linux (ubuntu) machine but
> am having all sorts of trouble following the readme, I get a ton of errors
> when trying to run the build.py. Mostly cannot find directory errors, but
> there are others thrown in for good measure.

Use prepackaged version of PIL if you can.
apt-get is your friend
Or for command line challenged, Synaptic.

The general instruction on installing:
https://help.ubuntu.com/6.10/ubuntu/desktopguide/C/add-applications.html

Waldemar

From fredrik at pythonware.com  Fri Jan 26 18:15:01 2007
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 26 Jan 2007 18:15:01 +0100
Subject: [Image-SIG] Newb help... can't install the PIL extension
	libraries
In-Reply-To: <45B907F2.7010104@noaa.gov>
References: <06C7A71EB82A7E4F8FBB65E016C15D07072A3B@kws-exc-01.kingswood.internal>
	<45B907F2.7010104@noaa.gov>
Message-ID: <epdcul$fdf$1@sea.gmane.org>

Christopher Barker wrote:

> oh, and are you sure there is not an existing Ubuntu package for PIL?

$ sudo apt-get install python-imaging

iirc, ubuntu-desktop uses it, so it's installed by default on most 
Ubuntu systems.

</F>


From wolfgangfellger at web.de  Sun Jan 28 21:24:04 2007
From: wolfgangfellger at web.de (Wolfgang Fellger)
Date: Sun, 28 Jan 2007 21:24:04 +0100
Subject: [Image-SIG] [BUG] PNG writer using 'bits' stores palette too big
Message-ID: <000001c7431a$4826c6e0$0509d452@wolff>

Sorry if this has been reported before, I was too lazy looking through all the archives.

When using the PNG writer with e.g. bits=4 it still stores a palette with full 256 entries into the resulting file.
This is not only a waste of space, it also makes e.g. GIMP silently convert it to a 8-bit palette image (and store it as such on save).

Solution: PNG writer should clip the palette as appropriate.


Keep up the good work!
-- 
Wolfgang Fellger

From robotissues at gmail.com  Fri Jan 26 16:43:13 2007
From: robotissues at gmail.com (Jason Van Anden)
Date: Fri, 26 Jan 2007 10:43:13 -0500
Subject: [Image-SIG] Splitting Color Stacked TIFFs Problem
In-Reply-To: <8eadd6950701251417r44baf177u54a35ee3aa110810@mail.gmail.com>
References: <8eadd6950701251417r44baf177u54a35ee3aa110810@mail.gmail.com>
Message-ID: <8eadd6950701260743o8ba0c64q601920960d842b15@mail.gmail.com>

Hello List,

A project I am developing in PIL splits stacked Tiffs (among other things).
It works super on B&W Tiffs, but chokes on color.

The error is "cannot identify image file".

My question is ... is this a limitation of PIL (Color Tiffs) or could
something be screwy about my files?  I am using the same program to capture
the images - just two different cameras.

Thank You,
Jason Van Anden
http://www.smileproject.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070126/0643a6e8/attachment.htm 

From fredrik at pythonware.com  Mon Jan 29 13:52:40 2007
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Mon, 29 Jan 2007 13:52:40 +0100
Subject: [Image-SIG] Splitting Color Stacked TIFFs Problem
References: <8eadd6950701251417r44baf177u54a35ee3aa110810@mail.gmail.com>
Message-ID: <epkqmo$8nr$1@sea.gmane.org>

Jason Van Anden wrote:

> A project I am developing in PIL splits stacked Tiffs (among other things).
> It works super on B&W Tiffs, but chokes on color.
>
> The error is "cannot identify image file".
>
> My question is ... is this a limitation of PIL (Color Tiffs) or could
> something be screwy about my files?  I am using the same program
> to capture the images - just two different cameras.

PIL's TIFF parser is pretty picky, so it's probably a bug or limitation in PIL.
Can you perhaps post a sample file somewhere, and mail me the URL ?

</F> 




From boesi.josi at gmx.net  Mon Jan 29 19:44:24 2007
From: boesi.josi at gmx.net (Alexander 'boesi' =?ISO-8859-1?Q?B=F6secke?=)
Date: Mon, 29 Jan 2007 19:44:24 +0100
Subject: [Image-SIG] video capture
In-Reply-To: <45B9273E.8010006@seehart.com>
References: <45B851A6.8050706@famillepinault.fr> <45B9273E.8010006@seehart.com>
Message-ID: <epliqp.2lo.1@boesi.MyFQDN.de>

Hi

Am 25.01.2007 22:55:10 schrieb Ken Seehart:

> Maybe what the world needs is a set of webcam drivers wrapped up into
> a single platform- independent PIL compatible python module, that
> could then be wrapped into a Kamelia component.  Anyone else
> interested in such a project?  I can't believe such a thing doesn't
> already exist.

Oh yeah I'm really interested in such a thing. At the moment I'm using
pyVideo [1], but this is a bit unstable and it doesn't offer a method to
set the resolution directly. Video Capture instead doesn't have an
useable preview. 

So I was thinking of coding a new "Video Capture Module", that can be
extended by drivers for different cameras and capture boards - I think
especially about the products from [2].

But I can only start the work, if I get paid for it. Well and at the
moment there is less money for the related projects. But when the money
flows again, I will truly start such a module and release it as open
source. And if somebody else will start something, I would contribute to
it ...


cu boesi

[1] http://www.geocities.com/rtrocca/python/
[2] http://www.ids-imaging.de/?&changeLang=e
-- 

...schlafen ist sowieso ungesund...
           .-==Police Academy I==-.

From djmuhlestein at gmail.com  Wed Jan 31 19:20:28 2007
From: djmuhlestein at gmail.com (Dennis Muhlestein)
Date: Wed, 31 Jan 2007 11:20:28 -0700
Subject: [Image-SIG] PIL potential bug
In-Reply-To: <1170267230.24405.4.camel@esteban>
References: <1170267230.24405.4.camel@esteban>
Message-ID: <eeed88270701311020i68c7ab30od6391e5b0b082270@mail.gmail.com>

Nope, for the time being, I'm simply not allowing PNG images.  I wish
I heard back from somebody about a bugfix or workaround, but nobody
responded.  I didn't have time to work on it then.

-Dennis

On 1/31/07, Florian Boucault <florian at boucault.ath.cx> wrote:
> Hello Dennis,
>
> I came accross the same problem as you with PIL that you described on
> the mailing list:
> http://www.mail-archive.com/image-sig at python.org/msg01153.html
>
> Did you manage to get it to work or did you use a workaround ?
>
> Thank you for your time,
>
> Florian Boucault
>
>


-- 
You can see what's happening at http://muhlesteins.com

From jason at smileproject.com  Wed Jan 31 18:41:42 2007
From: jason at smileproject.com (Jason Van Anden)
Date: Wed, 31 Jan 2007 12:41:42 -0500
Subject: [Image-SIG] Removing an Alpha Channel
Message-ID: <8eadd6950701310941l375e51dfj85d955b9d0ef01d8@mail.gmail.com>

I have written an app using PIL that uses alpha channels during processing,
but does not require them once the image is complete.  Is there an easy way
to remove existing alpha channels?  I suspect there is but the docs don't
make this obvious.

Thank You,
Jason Van Anden
www.smileproject.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20070131/e7ecb296/attachment.htm