From fredrik at pythonware.com  Wed Jan  2 10:49:11 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 02 Jan 2008 10:49:11 +0100
Subject: [Image-SIG] ImageFont Unicode Decode error
In-Reply-To: <200712261640.23347.donn.ingle@gmail.com>
References: <200712261640.23347.donn.ingle@gmail.com>
Message-ID: <flfmmo$3nm$1@ger.gmane.org>

Donn wrote:

>  Trying to open a ttf file with a non-standard character in it (u'\xe5') 
> causes an error, no matter what I pass to encoding:

>     self.font = _imagingft.getfont(file, size, index, encoding)
> UnicodeEncodeError: 'ascii' codec can't encode character u'\xe5' in position 
> 66: ordinal not in range(128)
> 
> Is there something I'm doing wrong, or is this a bug?

a glitch, at least.  the _imagingft driver appears to use Python's 
default encoding, rather than the filename encoding.

should be relatively easy to fix, I think, but in the meantime, you'll 
have to work around this by making a temporary copy of the file with a 
made-up name if this happens:

     try:
         font = ImageFont.truetype(filename, ...)
         altname = None
     except UnicodeEncodeError:
         # create a suitable alternative name (e.g. using the tempfile
         # module)
         altname = ...
         shutil.copy(filename, altname) # could use linking on Unix
         font = ImageFont.truetype(altname, ...)

     ... use the font object ...

     if altname:
         os.remove(altname)

</F>


From donn.ingle at gmail.com  Wed Jan  2 11:14:58 2008
From: donn.ingle at gmail.com (Donn)
Date: Wed, 2 Jan 2008 12:14:58 +0200
Subject: [Image-SIG] ImageFont Unicode Decode error
In-Reply-To: <flfmmo$3nm$1@ger.gmane.org>
References: <200712261640.23347.donn.ingle@gmail.com>
	<flfmmo$3nm$1@ger.gmane.org>
Message-ID: <200801021214.58612.donn.ingle@gmail.com>

> should be relatively easy to fix, I think, but in the meantime, you'll
> have to work around this by making a temporary copy of the file with a
> made-up name if this happens:
Thanks for the feedback and the idea.

\d

From fredrik at pythonware.com  Wed Jan  2 22:22:41 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 02 Jan 2008 22:22:41 +0100
Subject: [Image-SIG] ImageFont Unicode Decode error
In-Reply-To: <flfmmo$3nm$1@ger.gmane.org>
References: <200712261640.23347.donn.ingle@gmail.com>
	<flfmmo$3nm$1@ger.gmane.org>
Message-ID: <flgvb1$i02$1@ger.gmane.org>

I wrote:

> a glitch, at least.  the _imagingft driver appears to use Python's 
> default encoding, rather than the filename encoding.
> 
> should be relatively easy to fix, I think

For the record, here's a (preliminary) patch:

$ svn diff _imagingft.c
Index: _imagingft.c
===================================================================
--- _imagingft.c        (revision 3349)
+++ _imagingft.c        (working copy)
@@ -117,9 +118,17 @@
      static char* kwlist[] = {
          "filename", "size", "index", "encoding", NULL
      };
+
+#if defined(HAVE_UNICODE)
+    if (!PyArg_ParseTupleAndKeywords(args, kw, "eti|is", kwlist,
+                                     Py_FileSystemDefaultEncoding, 
&filename,
+                                     &size, &index, &encoding))
+        return NULL;
+#else
      if (!PyArg_ParseTupleAndKeywords(args, kw, "si|is", kwlist,
                                       &filename, &size, &index, &encoding))
          return NULL;
+#endif

      if (!library && FT_Init_FreeType(&library)) {
          PyErr_SetString(

</F>


From adam at volition-inc.com  Wed Jan  2 22:52:54 2008
From: adam at volition-inc.com (Adam Pletcher)
Date: Wed, 2 Jan 2008 15:52:54 -0600
Subject: [Image-SIG] buffer overrun loading some TGAs
Message-ID: <893A44FF792E904A97B7515CE341914601F9953C@volimxs01.thqinc.com>

Hello.  I searched the archives for this but came up empty.

I'm getting a "image buffer overrun error", when doing the first
operation on certain image files.  Specifically in \PIL\ImageFile.py,
line 207, PIL 1.1.6.

It appears limited to certain RGBA Targa images.  The odd part is, if
the operation is immediately run again it works fine with no errors.

For example:

   >>> img = Image.open(r'D:\temp\hud_sheet_04.tga')
   >>> img2 = img.convert('RGB')
   Traceback (most recent call last):
     File "<pyshell#13>", line 1, in <module>
       img2 = img.convert('RGB')
     File "C:\Python25\lib\site-packages\PIL\Image.py", line 653, in
convert
       self.load()
     File "C:\Python25\lib\site-packages\PIL\ImageFile.py", line 207, in
load
       raise IOError(error + " when reading image file")
   IOError: image buffer overrun error when reading image file
   >>> img2 = img.convert('RGB')
   >>>

Am I missing something?  Is there something non-standard about these
image files?  I'm guessing so, but I can't nail it down.  I've
successfully loaded them in in various programs with no issues.  It
appears I can get PIL to load them by re-saving them in certain
programs, even though no other apps give me load errors.

I posted one culprit image file in case someone is able to test/confirm
for me:
http://www.adamsarcade.com/misc/hud_sheet_04.tga

I can put a second operation attempt inside a try/except, but I'm
preparing this script for a talk I'm giving and I'd like to avoid that
kind of hackery if possible.

Any suggestions appreciated.

--
Adam Pletcher
Technical Art Director
Volition/THQ


From fredrik at pythonware.com  Wed Jan  2 23:29:55 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 02 Jan 2008 23:29:55 +0100
Subject: [Image-SIG] buffer overrun loading some TGAs
In-Reply-To: <893A44FF792E904A97B7515CE341914601F9953C@volimxs01.thqinc.com>
References: <893A44FF792E904A97B7515CE341914601F9953C@volimxs01.thqinc.com>
Message-ID: <flh394$vou$1@ger.gmane.org>

Adam Pletcher wrote:

> Am I missing something?  Is there something non-standard about these
> image files?  I'm guessing so, but I can't nail it down.  I've
> successfully loaded them in in various programs with no issues.  It
> appears I can get PIL to load them by re-saving them in certain
> programs, even though no other apps give me load errors.
> 
> I posted one culprit image file in case someone is able to test/confirm
> for me:
> http://www.adamsarcade.com/misc/hud_sheet_04.tga

I can repeat this with the development version.  It's an RLE-compressed 
targa file, so you've probably hit some edge case that PIL's decoder 
isn't handling properly.

 > It appears limited to certain RGBA Targa images.  The odd part is, if
 > the operation is immediately run again it works fine with no errors.

Have you checked that the resulting image contains the right data?

</F>


From adam at volition-inc.com  Wed Jan  2 23:45:56 2008
From: adam at volition-inc.com (Adam Pletcher)
Date: Wed, 2 Jan 2008 16:45:56 -0600
Subject: [Image-SIG] buffer overrun loading some TGAs
In-Reply-To: <flh394$vou$1@ger.gmane.org>
References: <893A44FF792E904A97B7515CE341914601F9953C@volimxs01.thqinc.com>
	<flh394$vou$1@ger.gmane.org>
Message-ID: <893A44FF792E904A97B7515CE341914601F99556@volimxs01.thqinc.com>

> I can repeat this with the development version.  It's an
RLE-compressed
> targa file, so you've probably hit some edge case that PIL's decoder
> isn't handling properly.

Ah, sweet.

> Have you checked that the resulting image contains the right data?

It doesn't, turns out.  It's the right size but the pixels are garbage.

- Adam

From fredrik at pythonware.com  Thu Jan  3 00:20:14 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Thu, 03 Jan 2008 00:20:14 +0100
Subject: [Image-SIG] buffer overrun loading some TGAs
In-Reply-To: <893A44FF792E904A97B7515CE341914601F99556@volimxs01.thqinc.com>
References: <893A44FF792E904A97B7515CE341914601F9953C@volimxs01.thqinc.com>	<flh394$vou$1@ger.gmane.org>
	<893A44FF792E904A97B7515CE341914601F99556@volimxs01.thqinc.com>
Message-ID: <flh67f$9la$1@ger.gmane.org>

Adam Pletcher wrote:

>> I can repeat this with the development version.  It's an
>> RLE-compressed targa file, so you've probably hit some edge
 >> case that PIL's decoder isn't handling properly.
> 
> Ah, sweet.

looks as if PIL's decode messes up for runs that spans multiple rows; 
this is obviously not very common (the decoder was written in 1997, and 
I'm pretty sure this hasn't been reported before), but appears to be 
perfectly legal according the Targa documentation.

time to dive into some ancient C code, it seems.

</F>


From carl at personnelware.com  Thu Jan  3 00:32:05 2008
From: carl at personnelware.com (Carl Karsten)
Date: Wed, 02 Jan 2008 17:32:05 -0600
Subject: [Image-SIG] less disk IO
Message-ID: <477C1EF5.9040903@personnelware.com>

I am trying to convert a pdf (data in string, not disk file) to a png.  This 
will be run in a django view on a server (PyCon's even) so it would be good if 
it didn't hit the disk as much.  Shouldn't I be able to use .tostring if I can 
use .save?

I gave up trying to convert the pdf.  The things that might have actually worked 
were not stable enough to be put onto the production server.   If you think you 
know a way, please code it first.  If it takes you longer than 10 min, or more 
than 100 words to describe the setup, it probably isn't going to get used :)

Also, I have been told about 10 times now that it should be easy.  I agree.  so 
far no one has proved it.

Thanks,
Carl K

def sample(ds):

     """ make a sample badge"""

     # this is some clunky code.
     # it bounces the data off the disk more times than I care to count
     # It would be nice if it was all done in memory,
     # but I couldn't figure out how.

     import PIL, Image
     import tempfile

     ds['attendeeID']='SAMPLE'
     pdf = mkpdf(ds)

     tempDir = tempfile.mkdtemp( prefix = "badgePreview_" )
     pdfname = path.join( tempDir, 'badge.pdf' )
     imgname = path.join( tempDir, 'badge.png' )

     open(pdfname,'wb').write(pdf)

     # convert pdf to some image format (uses ext to determine type)
     system('convert %s %s' % (pdfname, imgname))

     # crop the badge (pdf may be 8x11 page)
     Image.open(imgname).crop([0,0,300,235]).save(imgname)

     # I am sure there is a way to get the image out of im
     # so that I don't have to send it back to the disk.
     # im=Image.open(imgname)
     # im=im.crop([0,0,300,235])
     # img=im.tostring('raw') # The image x cannot be displayed, because it 
contains errors.
     # img=im.tostring('PNG') # encoder PNG not available
     # img=im.tostring('png') # encoder png not available

     img = open(imgname).read()
     remove(pdfname)
     remove(imgname)
     rmdir(tempDir)

     return img

From douglas at paradise.net.nz  Thu Jan  3 04:17:36 2008
From: douglas at paradise.net.nz (Douglas Bagnall)
Date: Thu, 03 Jan 2008 16:17:36 +1300
Subject: [Image-SIG] less disk IO
In-Reply-To: <477C1EF5.9040903@personnelware.com>
References: <477C1EF5.9040903@personnelware.com>
Message-ID: <477C53D0.3@paradise.net.nz>

Carl Karsten wrote:

> I am trying to convert a pdf (data in string, not disk file) to a png.  This 
> will be run in a django view on a server (PyCon's even) so it would be good if 
> it didn't hit the disk as much.  Shouldn't I be able to use .tostring if I can 
> use .save?

use cStringIO.

http://docs.python.org/lib/module-StringIO.html

It might not make much difference, because your OS should be caching the
file in memory anyway, and letting you read it back before it is
actually written to disk.

If you are sure it is a disk problem, you could set it up to write to a
ram file system (e.g. tmpfs on linux), which would help the PDF
conversion too.

Otherwise you could see if it possible to communicate with the convert
script using pipes, rather than system(). That will cut out a few reads
and writes.


Douglas

From carl at personnelware.com  Thu Jan  3 04:43:58 2008
From: carl at personnelware.com (Carl Karsten)
Date: Wed, 02 Jan 2008 21:43:58 -0600
Subject: [Image-SIG] less disk IO
In-Reply-To: <477C53D0.3@paradise.net.nz>
References: <477C1EF5.9040903@personnelware.com> <477C53D0.3@paradise.net.nz>
Message-ID: <477C59FE.60206@personnelware.com>

Douglas Bagnall wrote:
> Carl Karsten wrote:
> 
>> I am trying to convert a pdf (data in string, not disk file) to a png.  This 
>> will be run in a django view on a server (PyCon's even) so it would be good if 
>> it didn't hit the disk as much.  Shouldn't I be able to use .tostring if I can 
>> use .save?
> 
> use cStringIO.

I saw some mention of that in the IM docs.  any idea why .tostring doesn't work?

> 
> It might not make much difference, because your OS should be caching the
> file in memory anyway, and letting you read it back before it is
> actually written to disk.

busy server, so if I can easily make it better...

Also, keeping everything in memory means nothing to clean up when things crash. 
    which happens often enough, that if I can make it better...

> 
> If you are sure it is a disk problem, you could set it up to write to a
> ram file system (e.g. tmpfs on linux), which would help the PDF
> conversion too.

well, not sure.  just a feeling and some comments from an admin.
make it better... :)
> 
> Otherwise you could see if it possible to communicate with the convert
> script using pipes, rather than system(). That will cut out a few reads
> and writes.

I hear pipes aren't thread safe, and/or can deadlock the web server.  no time to 
  figure that out, so i am just avoiding it.

also, I just discovered http://packages.ubuntu.com/hardy/python/python-pythonmagick

but I can't find docs.  any suggestions?

(not sure if I will be able to install this package on the server it needs to be 
on, so I am going to try to spend less than 30 min on it ... might happen ;)

Carl K



> 
> 
> Douglas
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
> 
> 

From fredrik at pythonware.com  Thu Jan  3 11:22:14 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Thu, 03 Jan 2008 11:22:14 +0100
Subject: [Image-SIG] less disk IO
In-Reply-To: <477C59FE.60206@personnelware.com>
References: <477C1EF5.9040903@personnelware.com> <477C53D0.3@paradise.net.nz>
	<477C59FE.60206@personnelware.com>
Message-ID: <flid0n$66k$1@ger.gmane.org>

Carl Karsten wrote:

> I saw some mention of that in the IM docs.  any idea why .tostring doesn't work?

the tostring method returns pixel data as a packed array, not data in an 
image interchange file format.

(tostring/fromstring is a common data interchange protocol for array-
style objects in Python; see e.g. the array module and numerical python).

</F>


From carl at personnelware.com  Thu Jan  3 15:37:24 2008
From: carl at personnelware.com (Carl Karsten)
Date: Thu, 03 Jan 2008 08:37:24 -0600
Subject: [Image-SIG] less disk IO
In-Reply-To: <flid0n$66k$1@ger.gmane.org>
References: <477C1EF5.9040903@personnelware.com>
	<477C53D0.3@paradise.net.nz>	<477C59FE.60206@personnelware.com>
	<flid0n$66k$1@ger.gmane.org>
Message-ID: <477CF324.7020805@personnelware.com>

Fredrik Lundh wrote:
> Carl Karsten wrote:
> 
>> I saw some mention of that in the IM docs.  any idea why .tostring doesn't work?
> 
> the tostring method returns pixel data as a packed array, not data in an 
> image interchange file format.
> 
> (tostring/fromstring is a common data interchange protocol for array-
> style objects in Python; see e.g. the array module and numerical python).
> 

Ah!

Thank you.  now I don't feel so bad using StringIO.  I thought I just couldn't 
figure it out.

Carl K

From angelol at easyconnect.fr  Fri Jan  4 03:44:50 2008
From: angelol at easyconnect.fr (angelo)
Date: Fri, 04 Jan 2008 03:44:50 +0100
Subject: [Image-SIG] What's wrong for me ?
Message-ID: <1199414690.5194.5.camel@kubuntu1>

I'm searching ton send a message to the list with my new user
subscription (it's just a test message), but:
 
Your mail to 'Image-SIG' with the subject


   It's just a connect test to the list


Is being held until the list moderator can review it for approval.

The reason it is being held:

   Message has a suspicious header

Either the message will get posted to the list, or you will receive
notification of the moderator's decision.  If you would like to cancel
this posting, please visit the following URL:

...

a+

-- 
Venez faire un tour ici :

http://ekd.tolosano.info
http://monsitt.irruption.net
http://irruption.net/progdudim


From fredrik at pythonware.com  Fri Jan  4 14:54:45 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 04 Jan 2008 14:54:45 +0100
Subject: [Image-SIG] What's wrong for me ?
In-Reply-To: <1199414690.5194.5.camel@kubuntu1>
References: <1199414690.5194.5.camel@kubuntu1>
Message-ID: <flldr6$ebr$1@ger.gmane.org>

angelo wrote:

> I'm searching ton send a message to the list with my new user
> subscription (it's just a test message), but:

Please don't send test messages to the list; if you have something
to ask, just post the question.  If you don't get a reply, check the 
archives (mail.python.org) after a while.  If the message is there,
wait at least a few days before asking again.

> Your mail to 'Image-SIG' with the subject
> 
>    It's just a connect test to the list
> 
> Is being held until the list moderator can review it for approval.
> 
> The reason it is being held:
> 
>    Message has a suspicious header
> 
> Either the message will get posted to the list, or you will receive
> notification of the moderator's decision.  If you would like to cancel
> this posting, please visit the following URL:

I suspect that the mailing list is set up to catch things that look like 
administrative messages, test messages, and other "noise" messages.  The 
moderation queue is processed a few times per week.

</F>


From sedrik at gmail.com  Wed Jan  9 00:42:54 2008
From: sedrik at gmail.com (Andy McCurdy)
Date: Tue, 8 Jan 2008 15:42:54 -0800
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
Message-ID: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>

Hi all,

Just started using PIL.  I'm attempting to resize a .gif image to a
smaller size, keeping the height/width proportional.  The resize
works, but the resulting image is very grainy.  Resizing the same
image with ImageMagick's convert utility produced a far better quality
image.  I'd rather use PIL if possible since it has a much better API.
 The below code works fine with JPG files.  Am I missing something?

    from PIL import Image

    # original_content is a string of bytes representing the image.
    # the image is not stored locally on disk, so it's fetched from
the storage service as a string
    # and loaded up using a StringIO buffer
    # Have also tried loading the same .gif from a local disk directly
with Image.open('filename.gif'), but image after resize looked grainy
as well
    io = StringIO(original_content)
    pil_image = Image.open (io)

    # docs suggest that the default mode, 'P', won't use the ANTIALIAS
filter, so switch to RGB
    pil_image = pil_image.convert('RGB')


    # ... stuff happens here to figure out what the new width and
height should be

    # tuple representing the new size
    new_size = (new_width, new_height)

    # the actual resize...
    pil_image = pil_image.resize(new_size, Image.ANTIALIAS)

    # save the new image back out to a StringIO buffer -- has to be
sent back to the storage service
    io = StringIO()
    format = 'GIF'   # hardcoded in this example
    pil_image.save(io, format)



Thanks for any help possible.

From fredrik at pythonware.com  Wed Jan  9 09:28:41 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 09 Jan 2008 09:28:41 +0100
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>
Message-ID: <fm20jq$tjj$1@ger.gmane.org>

Andy McCurdy wrote:

> Just started using PIL.  I'm attempting to resize a .gif image to a
> smaller size, keeping the height/width proportional.  The resize
> works, but the resulting image is very grainy.

try converting the image to "RGB" first, and use resize(ANTIALIAS) on 
the RGB image before converting it back to P.

(if you resize a "P" image as is, you can only use the NEAREST 
resampling method.  this works reasonably well for certain kinds of 
"graphics", but less so for images that are dithered.  for some kinds of 
images, smoothing the RGB image before resizing it might also help.)

</F>


From spaceoutlet at gmail.com  Wed Jan  9 13:58:50 2008
From: spaceoutlet at gmail.com (Alex K)
Date: Wed, 9 Jan 2008 13:58:50 +0100
Subject: [Image-SIG] rounded corners
Message-ID: <e4a904790801090458k3a37a6d2vf8f45acf0976eafb@mail.gmail.com>

Hello,

Would anyone know how to generate thumbnails with rounded corners
using PIL? I'm also considering imagemagick if PIL turns out not to be
appropriate for the task.

Thank you so much,

Alex

From spe.stani.be at gmail.com  Wed Jan  9 14:43:02 2008
From: spe.stani.be at gmail.com (Stani)
Date: Wed, 09 Jan 2008 14:43:02 +0100
Subject: [Image-SIG] rounded corners
In-Reply-To: <e4a904790801090458k3a37a6d2vf8f45acf0976eafb@mail.gmail.com>
References: <e4a904790801090458k3a37a6d2vf8f45acf0976eafb@mail.gmail.com>
Message-ID: <4784CF66.8060203@gmail.com>

In PIL this is easy.

The following code preferably uses cache (just a simple python
dictionary) to avoid the overhead of generating the same transparency
masks. So typical use would be:

cache = {}
for image in images:
    thumbnail = image_round(image.thumbnail(size, filter), cache, radius)

Here is the code for rounded corners:

# Rounded corners
# (c) 2008 www.stani.be
# License: same as PIL

CROSS                   = _t('Cross')
ROUNDED                 = _t('Rounded')
SQUARE                  = _t('Square')

CORNERS                 = [ROUNDED,SQUARE,CROSS]
CORNER_ID               = 'rounded_corner_r%d_f%d'
CROSS_POS               = (CROSS,CROSS,CROSS,CROSS)
ROUNDED_POS             = (ROUNDED,ROUNDED,ROUNDED,ROUNDED)
ROUNDED_RECTANGLE_ID    = 'rounded_rectangle_r%d_f%d_s%s_p%s'

def round_image(image,cache={},radius=100,fill=255,pos=ROUNDED_POS,
        back_colour='#FFFFFF'):
    if image.mode != 'RGBA':
        image = image.convert('RGBA')
    mask    = create_rounded_rectangle(image.size,cache,radius,fill,pos)
    image.paste(Image.new('RGB',image.size,back_colour),(0,0),
        ImageChops.invert(mask))
    image.putalpha(mask)
    return image

def create_corner(radius=100,fill=255,factor=2):
    corner  = Image.new('L',(factor*radius,factor*radius),0)
    draw    = ImageDraw.Draw(corner)
    draw.pieslice((0,0,2*factor*radius,2*factor*radius),180,270,fill=fill)
    corner  = corner.resize((radius,radius),Image.ANTIALIAS)
    return corner

def create_rounded_rectangle(size=(600,400),cache={},radius=100,fill=255,
        pos=ROUNDED_POS):
    #rounded_rectangle
    im_x, im_y  = size
    rounded_rectangle_id    = ROUNDED_RECTANGLE_ID%(radius,fill,size,pos)
    if cache.has_key(rounded_rectangle_id):
        return cache[rounded_rectangle_id]
    else:
        #cross
        cross_id    = ROUNDED_RECTANGLE_ID%(radius,fill,size,CROSS_POS)
        if cache.has_key(cross_id):
            cross   = cache[cross_id]
        else:
            cross   = cache[cross_id]   = Image.new('L',size,0)
            draw    = ImageDraw.Draw(cross)
            draw.rectangle((radius,0,im_x-radius,im_y),fill=fill)
            draw.rectangle((0,radius,im_x,im_y-radius),fill=fill)
        if pos==CROSS_POS:
            return cross
        #corner
        corner_id   = CORNER_ID%(radius,fill)
        if cache.has_key(corner_id):
            corner  = cache[corner_id]
        else:
            corner  = cache[corner_id] = create_corner(radius,fill)
        #rounded rectangle
        rectangle   = Image.new('L',(radius,radius),255)
        rounded_rectangle   = cross.copy()
        for index, angle in enumerate(pos):
            if angle == CROSS:
                break
            if angle == ROUNDED:
                element = corner
            else:
                element = rectangle
            if index%2:
                x       = im_x-radius
                element = element.transpose(Image.FLIP_LEFT_RIGHT)
            else:
                x       = 0
            if index < 2:
                y       = 0
            else:
                y       = im_y-radius
                element = element.transpose(Image.FLIP_TOP_BOTTOM)
            rounded_rectangle.paste(element,(x,y))
        cache[rounded_rectangle_id] = rounded_rectangle
        return rounded_rectangle


Alex K schreef:
> Hello,
>
> Would anyone know how to generate thumbnails with rounded corners
> using PIL? I'm also considering imagemagick if PIL turns out not to be
> appropriate for the task.
>
> Thank you so much,
>
> Alex
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>   


From spe.stani.be at gmail.com  Wed Jan  9 14:45:31 2008
From: spe.stani.be at gmail.com (Stani)
Date: Wed, 09 Jan 2008 14:45:31 +0100
Subject: [Image-SIG] rounded corners
In-Reply-To: <4784CF66.8060203@gmail.com>
References: <e4a904790801090458k3a37a6d2vf8f45acf0976eafb@mail.gmail.com>
	<4784CF66.8060203@gmail.com>
Message-ID: <4784CFFB.5080904@gmail.com>

I forgot to take of the _t function, so the first lines should be:
CROSS                   = 'Cross'
ROUNDED                 = 'Rounded'
SQUARE                  = 'Square'

Stani schreef:
> In PIL this is easy.
>
> The following code preferably uses cache (just a simple python
> dictionary) to avoid the overhead of generating the same transparency
> masks. So typical use would be:
>
> cache = {}
> for image in images:
>    thumbnail = image_round(image.thumbnail(size, filter), cache, radius)
>
> Here is the code for rounded corners:
>
> # Rounded corners
> # (c) 2008 www.stani.be
> # License: same as PIL
>
> CROSS                   = _t('Cross')
> ROUNDED                 = _t('Rounded')
> SQUARE                  = _t('Square')
>
> CORNERS                 = [ROUNDED,SQUARE,CROSS]
> CORNER_ID               = 'rounded_corner_r%d_f%d'
> CROSS_POS               = (CROSS,CROSS,CROSS,CROSS)
> ROUNDED_POS             = (ROUNDED,ROUNDED,ROUNDED,ROUNDED)
> ROUNDED_RECTANGLE_ID    = 'rounded_rectangle_r%d_f%d_s%s_p%s'
>
> def round_image(image,cache={},radius=100,fill=255,pos=ROUNDED_POS,
>        back_colour='#FFFFFF'):
>    if image.mode != 'RGBA':
>        image = image.convert('RGBA')
>    mask    = create_rounded_rectangle(image.size,cache,radius,fill,pos)
>    image.paste(Image.new('RGB',image.size,back_colour),(0,0),
>        ImageChops.invert(mask))
>    image.putalpha(mask)
>    return image
>
> def create_corner(radius=100,fill=255,factor=2):
>    corner  = Image.new('L',(factor*radius,factor*radius),0)
>    draw    = ImageDraw.Draw(corner)
>    draw.pieslice((0,0,2*factor*radius,2*factor*radius),180,270,fill=fill)
>    corner  = corner.resize((radius,radius),Image.ANTIALIAS)
>    return corner
>
> def create_rounded_rectangle(size=(600,400),cache={},radius=100,fill=255,
>        pos=ROUNDED_POS):
>    #rounded_rectangle
>    im_x, im_y  = size
>    rounded_rectangle_id    = ROUNDED_RECTANGLE_ID%(radius,fill,size,pos)
>    if cache.has_key(rounded_rectangle_id):
>        return cache[rounded_rectangle_id]
>    else:
>        #cross
>        cross_id    = ROUNDED_RECTANGLE_ID%(radius,fill,size,CROSS_POS)
>        if cache.has_key(cross_id):
>            cross   = cache[cross_id]
>        else:
>            cross   = cache[cross_id]   = Image.new('L',size,0)
>            draw    = ImageDraw.Draw(cross)
>            draw.rectangle((radius,0,im_x-radius,im_y),fill=fill)
>            draw.rectangle((0,radius,im_x,im_y-radius),fill=fill)
>        if pos==CROSS_POS:
>            return cross
>        #corner
>        corner_id   = CORNER_ID%(radius,fill)
>        if cache.has_key(corner_id):
>            corner  = cache[corner_id]
>        else:
>            corner  = cache[corner_id] = create_corner(radius,fill)
>        #rounded rectangle
>        rectangle   = Image.new('L',(radius,radius),255)
>        rounded_rectangle   = cross.copy()
>        for index, angle in enumerate(pos):
>            if angle == CROSS:
>                break
>            if angle == ROUNDED:
>                element = corner
>            else:
>                element = rectangle
>            if index%2:
>                x       = im_x-radius
>                element = element.transpose(Image.FLIP_LEFT_RIGHT)
>            else:
>                x       = 0
>            if index < 2:
>                y       = 0
>            else:
>                y       = im_y-radius
>                element = element.transpose(Image.FLIP_TOP_BOTTOM)
>            rounded_rectangle.paste(element,(x,y))
>        cache[rounded_rectangle_id] = rounded_rectangle
>        return rounded_rectangle
>
>
> Alex K schreef:
>> Hello,
>>
>> Would anyone know how to generate thumbnails with rounded corners
>> using PIL? I'm also considering imagemagick if PIL turns out not to be
>> appropriate for the task.
>>
>> Thank you so much,
>>
>> Alex
>> _______________________________________________
>> Image-SIG maillist  -  Image-SIG at python.org
>> http://mail.python.org/mailman/listinfo/image-sig
>>   
>


From sedrik at gmail.com  Tue Jan  8 10:55:20 2008
From: sedrik at gmail.com (Andy McCurdy)
Date: Tue, 8 Jan 2008 01:55:20 -0800
Subject: [Image-SIG] Poor Image Quality When Resizing
Message-ID: <55bd76f30801080155m637a62bdic46dc1c21fb15c84@mail.gmail.com>

Hi all,

Just started using PIL.  I'm attempting to resize a .gif image to a smaller
size, keeping the height/width proportional.  The resize works, but the
resulting image is very grainy.  Resizing the same image with ImageMagick's
convert utility produced a far better quality image.  I'd rather use PIL if
possible since it has a much better API.  Am I missing something?

    from PIL import Image

    # original_content is a string of bytes representing the image.
    # the image is not stored locally on disk, so it's fetched from the
storage service as a string
    # and loaded up using a StringIO buffer
    # Have also tried loading the same .gif from a local disk directly with
Image.open('filename.gif'), but image after resize looked grainy as well
    io = StringIO(original_content)
    pil_image = Image.open(io)

    # docs suggest that the default mode, 'P', won't use the ANTIALIAS
filter, so switch to RGB
    pil_image = pil_image.convert('RGB')


    # ... stuff happens here to figure out what the new width and height
should be

    # tuple representing the new size
    new_size = (new_width, new_height)

    # the actual resize...
    pil_image = pil_image.resize(new_size, Image.ANTIALIAS)

    # save the new image back out to a StringIO buffer -- has to be sent
back to the storage service
    io = StringIO()
    format = 'GIF'   # hardcoded in this example
    pil_image.save(io, format)



Thanks for any help possible.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080108/05ebd312/attachment.htm 

From tiptip.barish at gmail.com  Tue Jan  8 20:54:13 2008
From: tiptip.barish at gmail.com (Varsha purohit)
Date: Tue, 8 Jan 2008 11:54:13 -0800
Subject: [Image-SIG]  Zooming any image
Message-ID: <9636adca0801081154u2af4da5al5479ad6c055df195@mail.gmail.com>

Hello All,
        In my program i have a gui where i am showing an image on the panel.
I want to implement zoom in and zoom out functions using wxpython and PIL. I
am tryin to find in PIL if there is any functionality like that.

Any help is appreciated.

-
Varsha
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080108/abc3a7d1/attachment.htm 

From fredrik at pythonware.com  Wed Jan  9 16:16:00 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 09 Jan 2008 16:16:00 +0100
Subject: [Image-SIG] Zooming any image
In-Reply-To: <9636adca0801081154u2af4da5al5479ad6c055df195@mail.gmail.com>
References: <9636adca0801081154u2af4da5al5479ad6c055df195@mail.gmail.com>
Message-ID: <fm2ofh$gn5$1@ger.gmane.org>

Varsha purohit wrote:

>         In my program i have a gui where i am showing an image on the 
> panel. I want to implement zoom in and zoom out functions using wxpython 
> and PIL. I am tryin to find in PIL if there is any functionality like that.

the transpose(EXTENT) method is often used for this:

    http://effbot.org/tag/PIL.Image.Image.transform

just pass in the size of the viewport and the area you want to display, 
given as a bounding box.

</F>


From andy at andymccurdy.com  Wed Jan  9 17:32:47 2008
From: andy at andymccurdy.com (Andy McCurdy)
Date: Wed, 9 Jan 2008 08:32:47 -0800
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <fm20jq$tjj$1@ger.gmane.org>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>
	<fm20jq$tjj$1@ger.gmane.org>
Message-ID: <55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>

Fredrik, thanks for the reply.  I am in fact converting to "RGB"
first... at least I think so.  Here's the code I'm using that produces
bad results:

from PIL import Image

# io is a file handle
pil_image = Image.open(io)

# convert to rgb
pil_image = pil_image.convert('RGB')

# new_width/new_height are calculated to retain the
# current height/width proportions
new_size = (new_width, new_height)

#resize the image
pil_image = pil_image.resize(new_size, Image.ANTIALIAS)

# save the image back to disk
pil_image.save(io, 'GIF')


On Jan 9, 2008 12:28 AM, Fredrik Lundh <fredrik at pythonware.com> wrote:
> Andy McCurdy wrote:
>
> > Just started using PIL.  I'm attempting to resize a .gif image to a
> > smaller size, keeping the height/width proportional.  The resize
> > works, but the resulting image is very grainy.
>
> try converting the image to "RGB" first, and use resize(ANTIALIAS) on
> the RGB image before converting it back to P.
>
> (if you resize a "P" image as is, you can only use the NEAREST
> resampling method.  this works reasonably well for certain kinds of
> "graphics", but less so for images that are dithered.  for some kinds of
> images, smoothing the RGB image before resizing it might also help.)
>
> </F>
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>

From fredrik at pythonware.com  Wed Jan  9 22:13:19 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 09 Jan 2008 22:13:19 +0100
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>	<fm20jq$tjj$1@ger.gmane.org>
	<55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>
Message-ID: <fm3ddg$188$1@ger.gmane.org>

Andy McCurdy wrote:

> Fredrik, thanks for the reply.  I am in fact converting to "RGB"
> first... at least I think so.  Here's the code I'm using that produces
> bad results:
> 
> from PIL import Image
> 
> # io is a file handle
> pil_image = Image.open(io)
> 
> # convert to rgb
> pil_image = pil_image.convert('RGB')
> 
> # new_width/new_height are calculated to retain the
> # current height/width proportions
> new_size = (new_width, new_height)
> 
> #resize the image
> pil_image = pil_image.resize(new_size, Image.ANTIALIAS)
> 
> # save the image back to disk
> pil_image.save(io, 'GIF')

is the grain you're talking about perhaps floyd-steinberg dithering?  do 
your images look better if you insert

    pil_image = pil_image.convert("P", dither=Image.NONE)

before you save the image?

</F>


From andy at andymccurdy.com  Wed Jan  9 22:59:41 2008
From: andy at andymccurdy.com (Andy McCurdy)
Date: Wed, 9 Jan 2008 13:59:41 -0800
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <fm3ddg$188$1@ger.gmane.org>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>
	<fm20jq$tjj$1@ger.gmane.org>
	<55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>
	<fm3ddg$188$1@ger.gmane.org>
Message-ID: <55bd76f30801091359q109b0c0bh2a61b14f13f4226c@mail.gmail.com>

Your suggestion didn't seem to fix the problem.  It will probably be
helpful for you to take a look at what I'm seeing.

Here's the original .gif file I'm testing with:
http://www.andymccurdy.com/original.gif

Here's the resized gif with the original code I pasted.  You'll see a
bunch of 'dots' throughout the image:
http://www.andymccurdy.com/resized_1.gif

Here's the image resized again with your suggestion of converting to
"P" mode and setting the dithering prior to saving:
http://www.andymccurdy.com/resized_2.gif

You'll see both of the resized images are pretty poor quality.

On Jan 9, 2008 1:13 PM, Fredrik Lundh <fredrik at pythonware.com> wrote:
> Andy McCurdy wrote:
>
> > Fredrik, thanks for the reply.  I am in fact converting to "RGB"
> > first... at least I think so.  Here's the code I'm using that produces
> > bad results:
> >
> > from PIL import Image
> >
> > # io is a file handle
> > pil_image = Image.open(io)
> >
> > # convert to rgb
> > pil_image = pil_image.convert('RGB')
> >
> > # new_width/new_height are calculated to retain the
> > # current height/width proportions
> > new_size = (new_width, new_height)
> >
> > #resize the image
> > pil_image = pil_image.resize(new_size, Image.ANTIALIAS)
> >
> > # save the image back to disk
> > pil_image.save(io, 'GIF')
>
> is the grain you're talking about perhaps floyd-steinberg dithering?  do
> your images look better if you insert
>
>     pil_image = pil_image.convert("P", dither=Image.NONE)
>
> before you save the image?
>
>
> </F>
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>

From douglas at paradise.net.nz  Wed Jan  9 23:57:12 2008
From: douglas at paradise.net.nz (douglas bagnall)
Date: Thu, 10 Jan 2008 11:57:12 +1300
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <55bd76f30801091359q109b0c0bh2a61b14f13f4226c@mail.gmail.com>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>
	<fm20jq$tjj$1@ger.gmane.org>
	<55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>
	<fm3ddg$188$1@ger.gmane.org>
	<55bd76f30801091359q109b0c0bh2a61b14f13f4226c@mail.gmail.com>
Message-ID: <d24e24ab0801091457p5b992eaey32808d1f52be7ef6@mail.gmail.com>

hi Andy

> Here's the resized gif with the original code I pasted.  You'll see a
> bunch of 'dots' throughout the image:
> http://www.andymccurdy.com/resized_1.gif

PIL uses the so-called "web palette" by default.  Your image uses
#fdfdfd for white, but the web palette uses #ffffff. So it dithers to
approximate #fdfdfd.  You probably really want the background to be
pure #ffffff white, because printers etc will also dither, but the
more immediate problem is the web palette. Try this:

pil_image = pil_image.convert("P", palette=Image.ADAPTIVE)

or

pil_image = pil_image.convert("P", dither=Image.NONE, palette=Image.ADAPTIVE)

These options are not well documented, but you can find them in the
PIL/Image.py source.

douglas

From andy at andymccurdy.com  Thu Jan 10 00:08:50 2008
From: andy at andymccurdy.com (Andy McCurdy)
Date: Wed, 9 Jan 2008 15:08:50 -0800
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <d24e24ab0801091457p5b992eaey32808d1f52be7ef6@mail.gmail.com>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>
	<fm20jq$tjj$1@ger.gmane.org>
	<55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>
	<fm3ddg$188$1@ger.gmane.org>
	<55bd76f30801091359q109b0c0bh2a61b14f13f4226c@mail.gmail.com>
	<d24e24ab0801091457p5b992eaey32808d1f52be7ef6@mail.gmail.com>
Message-ID: <55bd76f30801091508h20c0a0a6n83be9ffe312487b2@mail.gmail.com>

Hey Douglas,

I interpreted your response as a setting to change to make this
specific image resize better.  I'm using PIL to resize images that
users upload to my website.  Since I don't have control over the
media, I need generic settings that will be good for all users.  What
settings would you suggest to be best for all images, not just the
example I provided?

On Jan 9, 2008 2:57 PM, douglas bagnall <douglas at paradise.net.nz> wrote:
> hi Andy
>
> > Here's the resized gif with the original code I pasted.  You'll see a
> > bunch of 'dots' throughout the image:
> > http://www.andymccurdy.com/resized_1.gif
>
> PIL uses the so-called "web palette" by default.  Your image uses
> #fdfdfd for white, but the web palette uses #ffffff. So it dithers to
> approximate #fdfdfd.  You probably really want the background to be
> pure #ffffff white, because printers etc will also dither, but the
> more immediate problem is the web palette. Try this:
>
> pil_image = pil_image.convert("P", palette=Image.ADAPTIVE)
>
> or
>
> pil_image = pil_image.convert("P", dither=Image.NONE, palette=Image.ADAPTIVE)
>
> These options are not well documented, but you can find them in the
> PIL/Image.py source.
>
> douglas
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>

From douglas at paradise.net.nz  Thu Jan 10 00:34:07 2008
From: douglas at paradise.net.nz (douglas bagnall)
Date: Thu, 10 Jan 2008 12:34:07 +1300
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <55bd76f30801091508h20c0a0a6n83be9ffe312487b2@mail.gmail.com>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>
	<fm20jq$tjj$1@ger.gmane.org>
	<55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>
	<fm3ddg$188$1@ger.gmane.org>
	<55bd76f30801091359q109b0c0bh2a61b14f13f4226c@mail.gmail.com>
	<d24e24ab0801091457p5b992eaey32808d1f52be7ef6@mail.gmail.com>
	<55bd76f30801091508h20c0a0a6n83be9ffe312487b2@mail.gmail.com>
Message-ID: <d24e24ab0801091534j230ce51ai1ea92a0ecc95183d@mail.gmail.com>

Andy McCurdy <andy at andymccurdy.com> wrote:

> I interpreted your response as a setting to change to make this
> specific image resize better.  I'm using PIL to resize images that
> users upload to my website.  Since I don't have control over the
> media, I need generic settings that will be good for all users.  What
> settings would you suggest to be best for all images, not just the
> example I provided?

Using palette=Image.ADAPTIVE will be better for all images unless the
images are already in the web palette, or the viewers are from 1995
and have 256 colour displays.  Both of these are unlikely, and in any
case the adaptive system will work well enough for them. It will be
slower though.

douglas

From andy at andymccurdy.com  Thu Jan 10 00:45:16 2008
From: andy at andymccurdy.com (Andy McCurdy)
Date: Wed, 9 Jan 2008 15:45:16 -0800
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <d24e24ab0801091534j230ce51ai1ea92a0ecc95183d@mail.gmail.com>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>
	<fm20jq$tjj$1@ger.gmane.org>
	<55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>
	<fm3ddg$188$1@ger.gmane.org>
	<55bd76f30801091359q109b0c0bh2a61b14f13f4226c@mail.gmail.com>
	<d24e24ab0801091457p5b992eaey32808d1f52be7ef6@mail.gmail.com>
	<55bd76f30801091508h20c0a0a6n83be9ffe312487b2@mail.gmail.com>
	<d24e24ab0801091534j230ce51ai1ea92a0ecc95183d@mail.gmail.com>
Message-ID: <55bd76f30801091545j55c20738t3962ae7e9d016820@mail.gmail.com>

Thanks Douglas.  Just tested it out, and the quality has improved
dramatically on all GIF images I threw at it.

On Jan 9, 2008 3:34 PM, douglas bagnall <douglas at paradise.net.nz> wrote:
> Andy McCurdy <andy at andymccurdy.com> wrote:
>
> > I interpreted your response as a setting to change to make this
> > specific image resize better.  I'm using PIL to resize images that
> > users upload to my website.  Since I don't have control over the
> > media, I need generic settings that will be good for all users.  What
> > settings would you suggest to be best for all images, not just the
> > example I provided?
>
> Using palette=Image.ADAPTIVE will be better for all images unless the
> images are already in the web palette, or the viewers are from 1995
> and have 256 colour displays.  Both of these are unlikely, and in any
> case the adaptive system will work well enough for them. It will be
> slower though.
>
> douglas
>

From fredrik at pythonware.com  Thu Jan 10 08:45:18 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Thu, 10 Jan 2008 08:45:18 +0100
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <55bd76f30801091545j55c20738t3962ae7e9d016820@mail.gmail.com>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>	<fm20jq$tjj$1@ger.gmane.org>	<55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>	<fm3ddg$188$1@ger.gmane.org>	<55bd76f30801091359q109b0c0bh2a61b14f13f4226c@mail.gmail.com>	<d24e24ab0801091457p5b992eaey32808d1f52be7ef6@mail.gmail.com>	<55bd76f30801091508h20c0a0a6n83be9ffe312487b2@mail.gmail.com>	<d24e24ab0801091534j230ce51ai1ea92a0ecc95183d@mail.gmail.com>
	<55bd76f30801091545j55c20738t3962ae7e9d016820@mail.gmail.com>
Message-ID: <fm4ief$hn$1@ger.gmane.org>

Andy McCurdy wrote:

>> Using palette=Image.ADAPTIVE will be better for all images unless the
>> images are already in the web palette, or the viewers are from 1995
>> and have 256 colour displays.  Both of these are unlikely, and in any
>> case the adaptive system will work well enough for them. It will be
>> slower though.
 >
 > Thanks Douglas.  Just tested it out, and the quality has improved
 > dramatically on all GIF images I threw at it.

In addition to Douglas' excellent advice, you should probably also ask 
yourself if you really need to store the thumbnails as GIF images.

For the general case, I'd probably use JPEG for everything (if it's good 
enough for Google, etc).  If you have lots of non-photographic images in 
your database, you can use PNG for things that have 256 colors or less:

     if im.getcolors(256):
         # limited number of colors
         im = im.convert("P", palette=Image.ADAPTIVE, dither=Image.NONE)
         im.save(out, "PNG")
     else:
         im.save(out, "JPEG")

(getcolors returns None if there are more than the given number of 
colors in the image.  you can change 256 to something larger if you want 
to; 512 or 1024 probably works well in practice)

</F>


From Scott.Daniels at Acm.Org  Fri Jan 11 03:00:07 2008
From: Scott.Daniels at Acm.Org (Scott David Daniels)
Date: Thu, 10 Jan 2008 18:00:07 -0800
Subject: [Image-SIG] Poor Image Quality When Resizing a GIF
In-Reply-To: <fm4ief$hn$1@ger.gmane.org>
References: <55bd76f30801081542q625395f6m77ad0dda30d11dd7@mail.gmail.com>	<fm20jq$tjj$1@ger.gmane.org>	<55bd76f30801090832n768f08daoba298e20d7d7a81@mail.gmail.com>	<fm3ddg$188$1@ger.gmane.org>	<55bd76f30801091359q109b0c0bh2a61b14f13f4226c@mail.gmail.com>	<d24e24ab0801091457p5b992eaey32808d1f52be7ef6@mail.gmail.com>	<55bd76f30801091508h20c0a0a6n83be9ffe312487b2@mail.gmail.com>	<d24e24ab0801091534j230ce51ai1ea92a0ecc95183d@mail.gmail.com>	<55bd76f30801091545j55c20738t3962ae7e9d016820@mail.gmail.com>
	<fm4ief$hn$1@ger.gmane.org>
Message-ID: <fm6iad$se0$1@ger.gmane.org>

Fredrik Lundh wrote:
> For the general case, I'd probably use JPEG for everything (if it's good 
> enough for Google, etc).  If you have lots of non-photographic images in 
> your database, you can use PNG for things that have 256 colors or less:
>      if im.getcolors(256):
>          # limited number of colors
>          im = im.convert("P", palette=Image.ADAPTIVE, dither=Image.NONE)
>          im.save(out, "PNG")
>      else:
>          im.save(out, "JPEG")
PNG can cope with any number of colors (one of its advantages over GIF).
I suspect your advice is to determine the photo/non-photo property
by looking at the number of distinct colors.  Unfortunately, the naive
reader might infer that he couldn't put 300-color images in PNG.

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


From natures2ndsun at sbcglobal.net  Fri Jan 11 03:48:53 2008
From: natures2ndsun at sbcglobal.net (natures2ndsun)
Date: Thu, 10 Jan 2008 18:48:53 -0800
Subject: [Image-SIG] [PIL installation] missing ANSI C headers on Solaris =>
Message-ID: <000701c853fc$815b7de0$6072e545@BARKER>

III have Windows XP. recently I have been receiving warnings of Windows 95 - 98 ansi missing. My picture and fonts are distorted(grainy and large). I have downloaded whitesmoke recently and it is now refusing to display the  web page. My initial opening screen and icons are normal, but then it goes black, briefly, to return as large and grainy, Can you help with this problem?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080110/c35e3a9f/attachment.htm 

From tiptip.barish at gmail.com  Fri Jan 11 21:04:59 2008
From: tiptip.barish at gmail.com (Varsha purohit)
Date: Fri, 11 Jan 2008 12:04:59 -0800
Subject: [Image-SIG] [image-sig] controlling brightness of the image
Message-ID: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>

Hello All,
       I have written a program using ImageEnhance module of pil where i am
controlling brightness and contrast of the image but i  am unable to use
im.save method to save the file. My code is here :

import Image
import ImageFilter
import ImageEnhance

file2 = "xyz.jpg"
img2 = Image.open(file2)
ench = ImageEnhance.Brightness(img2)
ench1 = ench.filter(ImageFilter.SHARPEN)
ench.save("xyz1.jpg" + ext, "JPEG", quality=100)


It says Brightness doesnt have attributes save and filter ?

Than how can i save the enhanced image... any ideas...
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080111/b14dc5ef/attachment.htm 

From spe.stani.be at gmail.com  Fri Jan 11 23:16:56 2008
From: spe.stani.be at gmail.com (Stani)
Date: Fri, 11 Jan 2008 23:16:56 +0100
Subject: [Image-SIG] [image-sig] controlling brightness of the image
In-Reply-To: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>
References: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>
Message-ID: <4787EAD8.20103@gmail.com>

You have to save the image, so replace your last line with:
img2.save("xyz1.jpg",quality=100)

Stani

Varsha purohit schreef:
> Hello All,
>        I have written a program using ImageEnhance module of pil where i am
> controlling brightness and contrast of the image but i  am unable to use
> im.save method to save the file. My code is here :
>
> import Image
> import ImageFilter
> import ImageEnhance
>
> file2 = "xyz.jpg"
> img2 = Image.open(file2)
> ench = ImageEnhance.Brightness(img2)
> ench1 = ench.filter(ImageFilter.SHARPEN)
> ench.save("xyz1.jpg" + ext, "JPEG", quality=100)
>
>
> It says Brightness doesnt have attributes save and filter ?
>
> Than how can i save the enhanced image... any ideas...
>
>   
> ------------------------------------------------------------------------
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>   



From fredrik at pythonware.com  Sat Jan 12 09:29:28 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sat, 12 Jan 2008 09:29:28 +0100
Subject: [Image-SIG] [image-sig] controlling brightness of the image
In-Reply-To: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>
References: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>
Message-ID: <fm9tp8$880$1@ger.gmane.org>

Varsha purohit wrote:

>        I have written a program using ImageEnhance module of pil where i 
> am controlling brightness and contrast of the image but i  am unable to 
> use im.save method to save the file. My code is here :
> 
> import Image
> import ImageFilter
> import ImageEnhance
> 
> file2 = "xyz.jpg"
> img2 = Image.open(file2)
> ench = ImageEnhance.Brightness(img2)

this creates an enhancement object that can be used to create enhanced 
images.  you need to call the "enhance" method to actually get an image 
object.  see the documentation for details:

   http://www.pythonware.com/library/pil/handbook/imageenhance.htm

</F>


From tiptip.barish at gmail.com  Fri Jan 11 23:37:13 2008
From: tiptip.barish at gmail.com (Varsha purohit)
Date: Fri, 11 Jan 2008 14:37:13 -0800
Subject: [Image-SIG] [image-sig] controlling brightness of the image
In-Reply-To: <4787EAD8.20103@gmail.com>
References: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>
	<4787EAD8.20103@gmail.com>
Message-ID: <9636adca0801111437n330a9e03l50d7550af4fb235b@mail.gmail.com>

Hi Stani,
     But i have enhanced the image with variable name ench1. So will the
image be effected ??

thanks,

On Jan 11, 2008 2:16 PM, Stani <spe.stani.be at gmail.com> wrote:

> You have to save the image, so replace your last line with:
> img2.save("xyz1.jpg",quality=100)
>
> Stani
>
> Varsha purohit schreef:
> > Hello All,
> >        I have written a program using ImageEnhance module of pil where i
> am
> > controlling brightness and contrast of the image but i  am unable to use
> > im.save method to save the file. My code is here :
> >
> > import Image
> > import ImageFilter
> > import ImageEnhance
> >
> > file2 = "xyz.jpg"
> > img2 = Image.open(file2)
> > ench = ImageEnhance.Brightness(img2)
> > ench1 = ench.filter(ImageFilter.SHARPEN)
> > ench.save("xyz1.jpg" + ext, "JPEG", quality=100)
> >
> >
> > It says Brightness doesnt have attributes save and filter ?
> >
> > Than how can i save the enhanced image... any ideas...
> >
> >
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080111/f4c16b65/attachment.htm 

From spe.stani.be at gmail.com  Sat Jan 12 00:15:44 2008
From: spe.stani.be at gmail.com (Stani)
Date: Sat, 12 Jan 2008 00:15:44 +0100
Subject: [Image-SIG] [image-sig] controlling brightness of the image
In-Reply-To: <9636adca0801111437n330a9e03l50d7550af4fb235b@mail.gmail.com>
References: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>	
	<4787EAD8.20103@gmail.com>
	<9636adca0801111437n330a9e03l50d7550af4fb235b@mail.gmail.com>
Message-ID: <4787F8A0.1010102@gmail.com>

Sorry, I looked to quickly to your code. Your mistake is that you are mixin images and enhancers. For clarity I marked every image object with the suffix '_im':

import Image
import ImageFilter
import ImageEnhance

file2 = "xyz.jpg"
im = Image.open(file2)
enhancer = ImageEnhance.Brightness(im)
bright_im = enhancer.enhance(0.5) #any value you want
sharp_im = bright_im.filter(ImageFilter.SHARPEN)
sharp_im.save("xyz1.jpg",quality=100)

>From the documentation:

All enhancement classes implement a common interface, containing a single method:

enhancer.enhance(factor) => image

Returns an enhanced image. The factor is a floating point value controlling the enhancement. Factor 1.0 always returns a copy of the original image, lower factors mean less color (brightness, contrast, etc), and higher values more. There are no restrictions on this value.

http://www.pythonware.com/library/pil/handbook/imageenhance.htm



Varsha purohit schreef:
> Hi Stani,
>      But i have enhanced the image with variable name ench1. So will the
> image be effected ??
>
> thanks,
>
> On Jan 11, 2008 2:16 PM, Stani <spe.stani.be at gmail.com> wrote:
>
>   
>> You have to save the image, so replace your last line with:
>> img2.save("xyz1.jpg",quality=100)
>>
>> Stani
>>
>> Varsha purohit schreef:
>>     
>>> Hello All,
>>>        I have written a program using ImageEnhance module of pil where i
>>>       
>> am
>>     
>>> controlling brightness and contrast of the image but i  am unable to use
>>> im.save method to save the file. My code is here :
>>>
>>> import Image
>>> import ImageFilter
>>> import ImageEnhance
>>>
>>> file2 = "xyz.jpg"
>>> img2 = Image.open(file2)
>>> ench = ImageEnhance.Brightness(img2)
>>> ench1 = ench.filter(ImageFilter.SHARPEN)
>>> ench.save("xyz1.jpg" + ext, "JPEG", quality=100)
>>>
>>>
>>> It says Brightness doesnt have attributes save and filter ?
>>>
>>> Than how can i save the enhanced image... any ideas...
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> 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
>>
>>     
>
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080112/b542096b/attachment.htm 

From anewc2 at gmail.com  Sat Jan 12 00:24:53 2008
From: anewc2 at gmail.com (Amos Newcombe)
Date: Fri, 11 Jan 2008 18:24:53 -0500
Subject: [Image-SIG] [image-sig] controlling brightness of the image
In-Reply-To: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>
References: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>
Message-ID: <d789d67f0801111524h1e6e8c65ob22b9d923151000d@mail.gmail.com>

According to the docs, this should be the code:

import Image
import ImageFilter
import ImageEnhance

file2 = "xyz.jpg"
img2 = Image.open(file2)
ench = ImageEnhance.Brightness(img2)
img2 = ench.enhance(2) # change the factor 2 to fit your application
img2 = img2.filter(ImageFilter.SHARPEN)
img2.save("xyz1.jpg" + ext, "JPEG", quality=100)


It says Brightness doesnt have attributes save and filter ?
>

A brightness enhancer doesn't. An image does.

(I haven't checked your other steps. The save() call looks unnecessarily
complicated, but I don't know what code you've cut out.)

Amos
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080111/5acf1361/attachment.htm 

From fredrik at pythonware.com  Sat Jan 12 13:15:23 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sat, 12 Jan 2008 13:15:23 +0100
Subject: [Image-SIG] [image-sig] controlling brightness of the image
In-Reply-To: <d789d67f0801111524h1e6e8c65ob22b9d923151000d@mail.gmail.com>
References: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>
	<d789d67f0801111524h1e6e8c65ob22b9d923151000d@mail.gmail.com>
Message-ID: <fmab0r$l2v$1@ger.gmane.org>

Amos Newcombe wrote:

> > img2.save("xyz1.jpg" + ext, "JPEG", quality=100)
 >
> (I haven't checked your other steps. The save() call looks unnecessarily 
> complicated, but I don't know what code you've cut out.)

quality=100 is a bad idea, at least.  quoting myself from an earlier post:

   JPEG quality 100 is overkill, btw -- it completely disables JPEG's
   quantization stage, and "mainly of interest for experimental pur-
   poses", according to the JPEG library documentation, which
   continues:

     "Quality values above about 95 are NOT recommended for
     normal use; the compressed file size goes up dramatically
     for hardly any gain in output image quality."

As for suitable settings, the documentation recommends the following:

    "/.../ the quality setting should be between
    50 and 95; the default of 75 is often about right.  If
    you see defects at quality 75, then go up 5 or 10
    counts at a time until you are happy with the output
    image.  (The optimal setting will vary from one image
    to another.)

For the full excerpt from that documentation, see:

     http://mail.python.org/pipermail/image-sig/2002-July/001916.html

regards /F


From fredrik at pythonware.com  Sat Jan 12 13:29:48 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sat, 12 Jan 2008 13:29:48 +0100
Subject: [Image-SIG] [image-sig] controlling brightness of the image
In-Reply-To: <fmab0r$l2v$1@ger.gmane.org>
References: <9636adca0801111204p4e7d950andf345ac681361bd5@mail.gmail.com>	<d789d67f0801111524h1e6e8c65ob22b9d923151000d@mail.gmail.com>
	<fmab0r$l2v$1@ger.gmane.org>
Message-ID: <fmabrr$nd0$1@ger.gmane.org>

Fredrik Lundh wrote:

>      "Quality values above about 95 are NOT recommended for
>      normal use; the compressed file size goes up dramatically
>      for hardly any gain in output image quality."

to illustrate, here are the output sizes for the standard "LENNA" image, 
for some quality settings/formats:

  50 22833
  60 26253
  75 35292 (default)
  80 41162
  85 49294
  90 64280
  95 99870
100 219203
PNG 491272
RAW 786432 (uncompressed, w/o header)

</F>


From carl at personnelware.com  Sat Jan 12 17:36:06 2008
From: carl at personnelware.com (Carl Karsten)
Date: Sat, 12 Jan 2008 10:36:06 -0600
Subject: [Image-SIG] pdf to image
Message-ID: <4788EC76.7030804@personnelware.com>

For the PyCon registration web page, we would like to display a preview of what 
the persons badge will look like based on what they have entered.  (full name, 
nick name, extra lines in the fonts that will be used - helps make sure things 
will fit, etc.)

The badge gets created as a pdf=mkpdf(data set) - we can just give the person 
the pdf, but we are trying to put it on the web page with <img src..> so we need 
  to convert the pdf to an image.  I can't find a clean way of doing it.  I 
found no ghostscript bindings, and the 2 imagemagick bindings either screwed up 
stdio or segfaulted.  The best I could do is below, and it is causing this to 
show up in the error log:
close failed: [Errno 10] No child processes

and stack dumps:

 >     subprocess.call(gs_command)
 >
 >   File "/usr/local/lib/python2.5/subprocess.py", line 443, in call
 >     return Popen(*popenargs, **kwargs).wait()
 >
 >   File "/usr/local/lib/python2.5/subprocess.py", line 1110, in wait
 >     pid, sts = os.waitpid(self.pid, 0)
 >
 > OSError: [Errno 10] No child processes

here is the current code:

def sample(pdf):

     tempDir = tempfile.mkdtemp( prefix = "badgePreview_" )
     pdfname = path.join( tempDir, 'badge.pdf' )
     imgname = path.join( tempDir, 'badge.png' )

     open(pdfname,'wb').write(pdf)

     # convert pdf to some image format (uses ext to determine type)
     # system('convert %s %s' % (pdfname, imgname))
     # gs -sDEVICE=png16m -sOutputFile=x.png -dBATCH -dNOPAUSE badge.pdf
     gs_command = ['gs',
         '-q',
         '-sDEVICE=png16m',
         '-dBATCH', '-dNOPAUSE',
         '-sOutputFile=%s'%imgname,
         pdfname ]
     subprocess.call(gs_command)

     # crop the badge (pdf may be 8x11 page)
     # and write to a string that can be reutnred
     buffer = StringIO()
     Image.open(imgname).crop([0,0,300,235]).save(buffer, 'png')
     img = buffer.getvalue()

     # clean up
     buffer.close()
     remove(pdfname)
     remove(imgname)
     rmdir(tempDir)

     return img

So.... anyone have some nifty code?

Carl K

From fredrik at pythonware.com  Sat Jan 12 18:43:20 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sat, 12 Jan 2008 18:43:20 +0100
Subject: [Image-SIG] pdf to image
In-Reply-To: <4788EC76.7030804@personnelware.com>
References: <4788EC76.7030804@personnelware.com>
Message-ID: <fmau7n$d5f$1@ger.gmane.org>

Carl Karsten wrote:

> close failed: [Errno 10] No child processes
> 
> and stack dumps:
> 
>  >     subprocess.call(gs_command)
>  >
>  >   File "/usr/local/lib/python2.5/subprocess.py", line 443, in call
>  >     return Popen(*popenargs, **kwargs).wait()
>  >
>  >   File "/usr/local/lib/python2.5/subprocess.py", line 1110, in wait
>  >     pid, sts = os.waitpid(self.pid, 0)
>  >
>  > OSError: [Errno 10] No child processes

does the "gs" command behave properly if you run it from the shell?

</F>


From carl at personnelware.com  Sun Jan 13 16:53:01 2008
From: carl at personnelware.com (Carl Karsten)
Date: Sun, 13 Jan 2008 09:53:01 -0600
Subject: [Image-SIG] pdf to image
In-Reply-To: <fmau7n$d5f$1@ger.gmane.org>
References: <4788EC76.7030804@personnelware.com> <fmau7n$d5f$1@ger.gmane.org>
Message-ID: <478A33DD.4020206@personnelware.com>

Fredrik Lundh wrote:
> Carl Karsten wrote:
> 
>> close failed: [Errno 10] No child processes
>>
>> and stack dumps:
>>
>>  >     subprocess.call(gs_command)
>>  >
>>  >   File "/usr/local/lib/python2.5/subprocess.py", line 443, in call
>>  >     return Popen(*popenargs, **kwargs).wait()
>>  >
>>  >   File "/usr/local/lib/python2.5/subprocess.py", line 1110, in wait
>>  >     pid, sts = os.waitpid(self.pid, 0)
>>  >
>>  > OSError: [Errno 10] No child processes
> 
> does the "gs" command behave properly if you run it from the shell?
> 

Yes.

It even behaves most of the time when run from subprocess.call(gs_command).

Carl K

From fredrik at pythonware.com  Sun Jan 13 17:10:51 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sun, 13 Jan 2008 17:10:51 +0100
Subject: [Image-SIG] pdf to image
In-Reply-To: <478A33DD.4020206@personnelware.com>
References: <4788EC76.7030804@personnelware.com> <fmau7n$d5f$1@ger.gmane.org>
	<478A33DD.4020206@personnelware.com>
Message-ID: <fmdd6c$bfj$1@ger.gmane.org>

Carl Karsten wrote:

>>>  > OSError: [Errno 10] No child processes
 >>
>> does the "gs" command behave properly if you run it from the shell?
> 
> Yes.
> 
> It even behaves most of the time when run from subprocess.call(gs_command).

could it be this bug?

     http://bugs.python.org/issue1731717

if so, I guess switching to a plain old os.system would help.

quick fix: replace

     subprocess.call(gs_command)

with

     os.system(subprocess.list2cmdline(gs_command))

</F>


From Chris.Barker at noaa.gov  Mon Jan 14 20:44:38 2008
From: Chris.Barker at noaa.gov (Christopher Barker)
Date: Mon, 14 Jan 2008 11:44:38 -0800
Subject: [Image-SIG] pdf to image
In-Reply-To: <4788EC76.7030804@personnelware.com>
References: <4788EC76.7030804@personnelware.com>
Message-ID: <478BBBA6.8020102@noaa.gov>

OK, a bit (or a lot!), but once someone else brought it up---

Carl Karsten wrote:
>      # gs -sDEVICE=png16m -sOutputFile=x.png -dBATCH -dNOPAUSE badge.pdf

I've used similar code. However, in my case, we're thumbnailing 
arbitrary pdfs, which could be any size -- usually they are US-letter or 
A4, but there can be poster-sized ones there too. So my question is:

How to I tell ghostscript to make the resulting image a given size (say 
100x100 or less), or how do I use ghostscript to tell me the size of the 
pdf at hand, and then I could calculate the DPI myself? Any ideas?

To the OP: you don't seem to be specifying the DPI that you want. The 
default may be OK, but  you may want to be explicit:

-r150

gives you 150dpi.

-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 noahspurrier at gmail.com  Wed Jan 16 23:54:34 2008
From: noahspurrier at gmail.com (Noah.org)
Date: Wed, 16 Jan 2008 14:54:34 -0800
Subject: [Image-SIG] RMS difference patch
Message-ID: <c131f9f70801161454w2d559dabnea1b22c4047ff1f2@mail.gmail.com>

I have a patch that adds an RMS Difference method to the PIL.
Let me know if there is any interest in this. I originally wrote this to
help implement a video stabilization filter. You could also use
this to quantify how much a lossy compression algorithm distorts
an image relative to some other compression algorithm. I thought
others might find RMS Difference useful.

    http://www.noah.org/wiki/PIL_patches

This is a patch against the Python Imaging Library Imaging-1.1.6.
This adds the method ImageChops.difference_rms(im1, im2) which
returns a float of the RMS difference between the two given
images, im1 and im2. This is about 10 times faster than doing it
in Python. The RMS Difference can be used to detect and compare
changes in images (say, for motion detection) or to compare the
effect of different image operations. If the two images are
exactly equal then difference_rms() will return 0.0. As the
images diverge the RMS value returned will increase. It will
always be a positive number. In general, the greater the
difference between the two images then the greater the RMS
difference will be.

In video stabilization I take two frames of a video and shift one
of them around until the RMS Difference is minimized.

-- 
  Noah Spurrier | email: noah at noah.org

From douglas at paradise.net.nz  Thu Jan 17 01:27:16 2008
From: douglas at paradise.net.nz (Douglas Bagnall)
Date: Thu, 17 Jan 2008 13:27:16 +1300
Subject: [Image-SIG] RMS difference patch
In-Reply-To: <c131f9f70801161454w2d559dabnea1b22c4047ff1f2@mail.gmail.com>
References: <c131f9f70801161454w2d559dabnea1b22c4047ff1f2@mail.gmail.com>
Message-ID: <478EA0E4.7080907@paradise.net.nz>

hi Noah

> I have a patch that adds an RMS Difference method to the PIL.
> Let me know if there is any interest in this. I originally wrote this to
> help implement a video stabilization filter. You could also use
> this to quantify how much a lossy compression algorithm distorts
> an image relative to some other compression algorithm. I thought
> others might find RMS Difference useful.
> 
>     http://www.noah.org/wiki/PIL_patches
> 
> This is a patch against the Python Imaging Library Imaging-1.1.6.
> This adds the method ImageChops.difference_rms(im1, im2) which
> returns a float of the RMS difference between the two given
> images, im1 and im2. This is about 10 times faster than doing it
> in Python. The RMS Difference can be used to detect and compare
> changes in images (say, for motion detection) or to compare the
> effect of different image operations. If the two images are
> exactly equal then difference_rms() will return 0.0. As the
> images diverge the RMS value returned will increase. It will
> always be a positive number. In general, the greater the
> difference between the two images then the greater the RMS
> difference will be.

Very good. It is something I could use.

But there may be trouble with integer overflow. You are summing squared
differences in a signed long, which wraps at 2**31. And:

>>> 2**31 / (255 * 255)
33025L
>>> 33025 ** 0.5
181.72781845386248

So won't comparing black and white 182 x 182 images will wrap the
integer into a negative value and cause a floating point error in the
square root. Slightly larger pictures will wrap it right round to 0, and
  just give you a false answer.

Comparing pure black and white might seem capricious, but with larger
images the differnce doesn't need to be so great to overflow. For
example, with a 1024x1024 image:

>>> 2 ** (31-20) ** 0.5
9.9633077752473778

A pixel difference of 10 across the 1 megapixel image will wrap the sum
and crash.

So what can be done? Using an unsigned long would add a little bit of
room and take away the possibility of a negative sum, but it isn't much
better.

The other ways would be to use a long long (uint64_t) or double to sum
in, both of which might be a bit slower. I think doubles can sum
integral values up to around 2**52 (or 53?). That means your black and
white images can be quite big:

>>> ((2**52) / (255*255)) ** 0.5
263172.01568555878

and the error is less painful (because what happens at that limit is
just that the limit + 1 equals the limit).

I haven't tested your code or anything, but this is a kind of problem I
have met in the past. I'd recommend you try doubles or longer ints and
see whether the slow down is tolerable. I suspect it will be, and that
you could offset it anyway by taking the "++ count" out of the loop and
using instead:

count = imIn1->xsize * imIn1->ysize;


regards,

Douglas

From douglas at paradise.net.nz  Thu Jan 17 01:37:01 2008
From: douglas at paradise.net.nz (Douglas Bagnall)
Date: Thu, 17 Jan 2008 13:37:01 +1300
Subject: [Image-SIG] RMS difference patch
In-Reply-To: <478EA0E4.7080907@paradise.net.nz>
References: <c131f9f70801161454w2d559dabnea1b22c4047ff1f2@mail.gmail.com>
	<478EA0E4.7080907@paradise.net.nz>
Message-ID: <478EA32D.901@paradise.net.nz>

I wrote:


> 
>>>> 2 ** (31-20) ** 0.5
> 9.9633077752473778
> 
> A pixel difference of 10 across the 1 megapixel image will wrap the sum
> and crash.


But now I've read it and it looks like it should have been

>>> (2 ** (31-20)) ** 0.5
45.254833995939045

So, not so bad, but nonetheless a problem.

douglas

From fredrik at pythonware.com  Thu Jan 17 11:14:54 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Thu, 17 Jan 2008 11:14:54 +0100
Subject: [Image-SIG] RMS difference patch
In-Reply-To: <c131f9f70801161454w2d559dabnea1b22c4047ff1f2@mail.gmail.com>
References: <c131f9f70801161454w2d559dabnea1b22c4047ff1f2@mail.gmail.com>
Message-ID: <fmn9qu$220$1@ger.gmane.org>

Noah.org wrote:

> I have a patch that adds an RMS Difference method to the PIL.
> Let me know if there is any interest in this. I originally wrote this to
> help implement a video stabilization filter. You could also use
> this to quantify how much a lossy compression algorithm distorts
> an image relative to some other compression algorithm. I thought
> others might find RMS Difference useful.
> 
>     http://www.noah.org/wiki/PIL_patches
> 
> This is a patch against the Python Imaging Library Imaging-1.1.6.
> This adds the method ImageChops.difference_rms(im1, im2) which
> returns a float of the RMS difference between the two given
> images, im1 and im2.

isn't this equivalent to

     ImageStat.Stat(ImageChops.difference(i, im)).rms

?

(I'm busy with other stuff, so I might be missing something here).

</F>


From imgrey at gmail.com  Sat Jan 19 11:31:35 2008
From: imgrey at gmail.com (Vitaliyi)
Date: Sat, 19 Jan 2008 12:31:35 +0200
Subject: [Image-SIG] format is None after convert('RGB')
Message-ID: <3aac340801190231r46a4809fs4d36715e1cf1eda6@mail.gmail.com>

Hi

"""
from PIL import Image
img = Image.open('/tmp/tst.png')

#after following conversion format is unknown. is it correct ?
if img.mode != 'RGB':
    img = img.convert('RGB')

P.S. I'm using python-imaging 1.1.6-1 (from debian)

Thanks

From fredrik at pythonware.com  Sat Jan 19 12:16:51 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sat, 19 Jan 2008 12:16:51 +0100
Subject: [Image-SIG] format is None after convert('RGB')
In-Reply-To: <3aac340801190231r46a4809fs4d36715e1cf1eda6@mail.gmail.com>
References: <3aac340801190231r46a4809fs4d36715e1cf1eda6@mail.gmail.com>
Message-ID: <fmsm73$4q3$1@ger.gmane.org>

Vitaliyi wrote:

> from PIL import Image
> img = Image.open('/tmp/tst.png')
> 
> #after following conversion format is unknown. is it correct ?
> if img.mode != 'RGB':
>     img = img.convert('RGB')
> 
> P.S. I'm using python-imaging 1.1.6-1 (from debian)

the "format" attribute only reflects the immediate source.  like most 
PIL operations, "convert" returns a *new* image object.

</F>


From drever at in.tum.de  Mon Jan 21 16:56:59 2008
From: drever at in.tum.de (drever)
Date: Mon, 21 Jan 2008 16:56:59 +0100
Subject: [Image-SIG] aggdraw
Message-ID: <4794C0CB.4030304@in.tum.de>

Hello!

I'm using PIL to load and save images. To draw images I use ImageDraw, 
but I am not content with the drawing quality. aggdraw uses agg which is 
a high quality drawing engine. I could not install it, so is there an 
alternative or can anybody help?

I tried to install it on ubuntu 7.10 with python 2.5.1. After building 
the package the ./selftest.py crashed with a segfault.

When I import aggdraw to python, I can use it with Image without 
crashes, but I don't see anything on the Image.

Thanks for any advice!

Johannes

From coen at reservoir.nl  Mon Jan 21 17:45:39 2008
From: coen at reservoir.nl (Coen van der Kamp)
Date: Mon, 21 Jan 2008 17:45:39 +0100
Subject: [Image-SIG] aggdraw
In-Reply-To: <4794C0CB.4030304@in.tum.de>
References: <4794C0CB.4030304@in.tum.de>
Message-ID: <4794CC33.9040304@reservoir.nl>

drever wrote:
> Hello!
>
> I'm using PIL to load and save images. To draw images I use ImageDraw, 
> but I am not content with the drawing quality. aggdraw uses agg which is 
> a high quality drawing engine. I could not install it, so is there an 
> alternative or can anybody help?
>
> I tried to install it on ubuntu 7.10 with python 2.5.1. After building 
> the package the ./selftest.py crashed with a segfault.
>
> When I import aggdraw to python, I can use it with Image without 
> crashes, but I don't see anything on the Image.
>
> Thanks for any advice!
>
> Johannes
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>
>   
Hello Drever,

You don't have errors when importing aggdraw and Image? Than this has to 
work:

import aggdraw, Image
im = Image.new("RGB", (100, 100), ("white"))
draw = aggdraw.Draw(im)
pen = aggdraw.Pen('black', width=5, opacity=255)
draw.ellipse((10,10,90,90), pen)
draw.flush() # Without the flush you'll get an empty image!
# im.show() # Has compression
im.save("/path/to/your/folder/test.png", "png") # Clean aggdraw image

Hope it helps!
Coen

From drever at in.tum.de  Tue Jan 22 10:28:42 2008
From: drever at in.tum.de (drever)
Date: Tue, 22 Jan 2008 10:28:42 +0100
Subject: [Image-SIG] aggdraw
In-Reply-To: <4794CC33.9040304@reservoir.nl>
References: <4794C0CB.4030304@in.tum.de> <4794CC33.9040304@reservoir.nl>
Message-ID: <4795B74A.8050800@in.tum.de>

Hello Coen,

thank you, that helped! I didn't notice the flush statement.

Johannes

Coen van der Kamp wrote:
> drever wrote:
>> Hello!
>>
>> I'm using PIL to load and save images. To draw images I use 
>> ImageDraw, but I am not content with the drawing quality. aggdraw 
>> uses agg which is a high quality drawing engine. I could not install 
>> it, so is there an alternative or can anybody help?
>>
>> I tried to install it on ubuntu 7.10 with python 2.5.1. After 
>> building the package the ./selftest.py crashed with a segfault.
>>
>> When I import aggdraw to python, I can use it with Image without 
>> crashes, but I don't see anything on the Image.
>>
>> Thanks for any advice!
>>
>> Johannes
>> _______________________________________________
>> Image-SIG maillist  -  Image-SIG at python.org
>> http://mail.python.org/mailman/listinfo/image-sig
>>
>>   
> Hello Drever,
>
> You don't have errors when importing aggdraw and Image? Than this has 
> to work:
>
> import aggdraw, Image
> im = Image.new("RGB", (100, 100), ("white"))
> draw = aggdraw.Draw(im)
> pen = aggdraw.Pen('black', width=5, opacity=255)
> draw.ellipse((10,10,90,90), pen)
> draw.flush() # Without the flush you'll get an empty image!
> # im.show() # Has compression
> im.save("/path/to/your/folder/test.png", "png") # Clean aggdraw image
>
> Hope it helps!
> Coen


From sylvain.berger at alpha-vision.com  Thu Jan 17 01:58:29 2008
From: sylvain.berger at alpha-vision.com (Sylvain Berger)
Date: Wed, 16 Jan 2008 19:58:29 -0500
Subject: [Image-SIG] Maya iff support for PIL
Message-ID: <3E284826C76A844A9D14DB667E6AFE7F01BC178A@alpha-dc.alpha-vision.com>

Hi, I am looking for a way to add the Maya iff format to PIL.

 

Any help would be appreciated

 

Sylvain Berger | Technical Director | Alpha Vision

Alpha Vision, 420 Armand Frappier suite 250, Laval, Canada H7V 4B4

(450) 902-0222 ext.248

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080116/99756c11/attachment.htm 

From Steven.Watanabe at autodesk.com  Tue Jan 22 15:24:46 2008
From: Steven.Watanabe at autodesk.com (Steven Watanabe)
Date: Tue, 22 Jan 2008 06:24:46 -0800
Subject: [Image-SIG] PIL and HDR images
Message-ID: <800886C2A4A73E44BA0FE45152C023702503AD651A@ADSK-NAMSG-02.MGDADSK.autodesk.com>

PIL 1.1.6 out of the box does not appear to support HDR images, correct? Does anyone know if an HDR image plugin is available anywhere?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080122/900f0233/attachment.htm 

From adam at volition-inc.com  Wed Jan 23 00:01:08 2008
From: adam at volition-inc.com (Adam Pletcher)
Date: Tue, 22 Jan 2008 17:01:08 -0600
Subject: [Image-SIG] writing TGAs?
Message-ID: <893A44FF792E904A97B7515CE341914602250306@volimxs01.thqinc.com>

Does anyone know if writing .TGA files will be supported in a future
release?

That would sure help me out, at least.

--
Adam Pletcher
Technical Art Director
Volition/THQ


From pyprog05 at gmail.com  Wed Jan 23 00:11:07 2008
From: pyprog05 at gmail.com (PyProg PyProg)
Date: Wed, 23 Jan 2008 00:11:07 +0100
Subject: [Image-SIG] writing TGAs?
In-Reply-To: <893A44FF792E904A97B7515CE341914602250306@volimxs01.thqinc.com>
References: <893A44FF792E904A97B7515CE341914602250306@volimxs01.thqinc.com>
Message-ID: <b763fa190801221511y38ea6ale240bd85c8fcc365@mail.gmail.com>

2008/1/23, Adam Pletcher <adam at volition-inc.com>:
> Does anyone know if writing .TGA files will be supported in a future
> release?
>
> That would sure help me out, at least.

I'm also interested with this.

a+

-- 
http://ekd.tolosano.info

From gwidion at mpc.com.br  Wed Jan 23 00:20:41 2008
From: gwidion at mpc.com.br (Joao S. O. Bueno)
Date: Tue, 22 Jan 2008 20:20:41 -0300
Subject: [Image-SIG] writing TGAs?
In-Reply-To: <893A44FF792E904A97B7515CE341914602250306@volimxs01.thqinc.com>
References: <893A44FF792E904A97B7515CE341914602250306@volimxs01.thqinc.com>
Message-ID: <200801222020.41468.gwidion@mpc.com.br>

On Tuesday 22 January 2008 20:01, Adam Pletcher wrote:
> Does anyone know if writing .TGA files will be supported in a
> future release?
>
> That would sure help me out, at least.
>

Hi there.

I am a coder, and work as a freelancer. I don't  know  if TGA support 
is supposed to be added to PIL - I'd say that not since TGA is no 
longer a popular format.  

However, as we are dealing with Free Software,  you could hire someone 
to write this support for a fee.   It even be cheap - and therefore 
you'd have he feature you need, and donate it to the World.

Other options would be using other python-2d image libraries. I don't 
know if any besides  GIMP-Python can  write TGAs though, and that one 
must be run as  a GIMP plug-in (so, you need to fire up a GIMP 
proccess).


Regards,
	js
	-><-


> Adam Pletcher
> Technical Art Director
> Volition/THQ

From spe.stani.be at gmail.com  Wed Jan 23 13:28:44 2008
From: spe.stani.be at gmail.com (Stani)
Date: Wed, 23 Jan 2008 13:28:44 +0100
Subject: [Image-SIG] Problem with CMYK jpeg images
Message-ID: <479732FC.8020606@gmail.com>

Hi,
PIL distorts the colours of a collection of CMYK jpeg images, even when 
you just try:
im = Image.open(filename)
im.save(filename)

You can get an example image here:
http://launchpadlibrarian.net/11232536/Marino_Detail02.jpg

How can I fix this? Is there a workaround? Or is this a bug in PIL?

Thanks for any help,
Stani

From smmoore at tiscali.co.za  Thu Jan 24 09:47:45 2008
From: smmoore at tiscali.co.za (Samuel Moore)
Date: Thu, 24 Jan 2008 10:47:45 +0200
Subject: [Image-SIG] PIL - How do I resize an image while keeping the IPTC
	and EXIF info?
Message-ID: <000601c85e65$cad3dfd0$0300a8c0@blindlemonsam>

Hi,
I'm using the Python Imaging Library (PIL v1.1.6) to batch resize images. Problem is, it erases the IPTC and EXIF data for the resized file. I want the data to stay. Help??
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080124/2d735149/attachment.htm 

From wrybread at gmail.com  Sun Jan 27 01:58:42 2008
From: wrybread at gmail.com (Alec Bennett)
Date: Sat, 26 Jan 2008 16:58:42 -0800
Subject: [Image-SIG] PIL - How do I resize an image while keeping the
	IPTC and EXIF info?
In-Reply-To: <000601c85e65$cad3dfd0$0300a8c0@blindlemonsam>
References: <000601c85e65$cad3dfd0$0300a8c0@blindlemonsam>
Message-ID: <cab7082b0801261658t2fff09d6k901b4cd481e88c8f@mail.gmail.com>

I don't know if the PIL has a way to preserve it, but if not you could
of course read the data before resizing then write the data to the new
image.

http://simonwillison.net/2003/Nov/13/exif/






2008/1/24 Samuel Moore <smmoore at tiscali.co.za>:
>
>
> Hi,
> I'm using the Python Imaging Library (PIL v1.1.6) to batch resize images.
> Problem is, it erases the IPTC and EXIF data for the resized file. I want
> the data to stay. Help??
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>
>

From fredrik at pythonware.com  Mon Jan 28 23:07:23 2008
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Mon, 28 Jan 2008 23:07:23 +0100
Subject: [Image-SIG] PIL - How do I resize an image while keeping the
 IPTC and EXIF info?
In-Reply-To: <000601c85e65$cad3dfd0$0300a8c0@blindlemonsam>
References: <000601c85e65$cad3dfd0$0300a8c0@blindlemonsam>
Message-ID: <fnljmr$htk$1@ger.gmane.org>

Samuel Moore wrote:

> I'm using the Python Imaging Library (PIL v1.1.6) to batch resize 
> images. Problem is, it erases the IPTC and EXIF data for the resized 
> file. I want the data to stay. Help??

PIL cannot save IPTC/EXIF data, so there's no way to do this "out of the 
box".

(patches are welcome, as usual).

</F>


From ndo at gmx.de  Tue Jan 29 13:21:30 2008
From: ndo at gmx.de (Nando)
Date: Tue, 29 Jan 2008 13:21:30 +0100
Subject: [Image-SIG] Maya iff support for PIL
In-Reply-To: <3E284826C76A844A9D14DB667E6AFE7F01BC178A@alpha-dc.alpha-vision.com>
References: <3E284826C76A844A9D14DB667E6AFE7F01BC178A@alpha-dc.alpha-vision.com>
Message-ID: <479F1A4A.5000405@gmx.de>

wow, Maya iff support would be extraordinary cool.
Maybe this could help:
http://www.stdout.org/~luke/iff2tga/

Cheers, Nando.

Sylvain Berger wrote:
> Hi, I am looking for a way to add the Maya iff format to PIL.
> 
>  
> 
> Any help would be appreciated
> 
>  
> 
> *Sylvain Berger** **| Technical Director | Alpha Vision*
> 
> Alpha Vision, 420 Armand Frappier suite 250, Laval, Canada H7V 4B4
> 
> (450) 902-0222 ext.248
> 
>  
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig



From kwangyul.seo at gmail.com  Tue Jan 29 04:15:03 2008
From: kwangyul.seo at gmail.com (KwangYul Seo)
Date: Tue, 29 Jan 2008 12:15:03 +0900
Subject: [Image-SIG] Conversion to ARGB format.
Message-ID: <7beb12420801281915w12484dd5n570ef9819709f4f@mail.gmail.com>

Hi,

Is it possible to convert a RGB image to ARGB format?  The following code
did not work.

import PIL.Image
im = PIL.Imageim.new("RGB", (100, 100))
im.convert("ARGB")


Thanks,
Kwang Yul Seo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080129/f730598c/attachment.htm 

From spe.stani.be at gmail.com  Wed Jan 30 03:50:00 2008
From: spe.stani.be at gmail.com (SPE Stani's Python Editor)
Date: Wed, 30 Jan 2008 03:50:00 +0100
Subject: [Image-SIG] Problem with CMYK jpeg images
In-Reply-To: <479732FC.8020606@gmail.com>
References: <479732FC.8020606@gmail.com>
Message-ID: <2078a7ad0801291850v204950e8ye493ce9e84fa9f7a@mail.gmail.com>

After digging in the archives, I found a solution in this thread:
http://www.mail-archive.com/image-sig at python.org/msg01213.html

Is there any reason why this was rejected?

Stani

On Jan 23, 2008 1:28 PM, Stani <spe.stani.be at gmail.com> wrote:
> Hi,
> PIL distorts the colours of a collection of CMYK jpeg images, even when
> you just try:
> im = Image.open(filename)
> im.save(filename)
>
> You can get an example image here:
> http://launchpadlibrarian.net/11232536/Marino_Detail02.jpg
>
> How can I fix this? Is there a workaround? Or is this a bug in PIL?
>
> Thanks for any help,
> Stani
>



-- 
http://pythonide.stani.be

From rsenra at acm.org  Wed Jan 30 02:54:39 2008
From: rsenra at acm.org (Rodrigo Dias Arruda Senra)
Date: Tue, 29 Jan 2008 23:54:39 -0200
Subject: [Image-SIG] Conversion to ARGB format.
In-Reply-To: <7beb12420801281915w12484dd5n570ef9819709f4f@mail.gmail.com>
References: <7beb12420801281915w12484dd5n570ef9819709f4f@mail.gmail.com>
Message-ID: <20080129235439.354e7158@Neo>

[ KwangYul Seo ]:
---------------------
>Hi,
>
>Is it possible to convert a RGB image to ARGB format?  The following
>code did not work.
>
>import PIL.Image
>im = PIL.Imageim.new("RGB", (100, 100))
>im.convert("ARGB")

Hi, 
 have you tried RGBA instead of ARGB ?

cheers,
Senra

From wrybread at gmail.com  Wed Jan 30 07:34:22 2008
From: wrybread at gmail.com (Alec Bennett)
Date: Wed, 30 Jan 2008 07:34:22 +0100
Subject: [Image-SIG] Conversion to ARGB format.
In-Reply-To: <7beb12420801281915w12484dd5n570ef9819709f4f@mail.gmail.com>
References: <7beb12420801281915w12484dd5n570ef9819709f4f@mail.gmail.com>
Message-ID: <cab7082b0801292234o1c68bfd6j1beabd452bc4cf20@mail.gmail.com>

im = Image.open("whatever.jpg").convert('RGBA')



2008/1/29 KwangYul Seo <kwangyul.seo at gmail.com>:
> Hi,
>
> Is it possible to convert a RGB image to ARGB format?  The following code
> did not work.
>
> import PIL.Image
> im = PIL.Imageim.new("RGB", (100, 100))
> im.convert("ARGB")
>
>
> Thanks,
>  Kwang Yul Seo
>
>
> _______________________________________________
> Image-SIG maillist  -  Image-SIG at python.org
> http://mail.python.org/mailman/listinfo/image-sig
>
>

From ned at nedbatchelder.com  Thu Jan 31 00:24:38 2008
From: ned at nedbatchelder.com (Ned Batchelder)
Date: Wed, 30 Jan 2008 18:24:38 -0500
Subject: [Image-SIG] Truly transparent text with PIL
Message-ID: <47A10736.1050209@nedbatchelder.com>

I know this is a subject that comes up occasionally on this list, so I 
thought readers would be interested in how to make truly transparent 
text with PIL: 
http://nedbatchelder.com/blog/200801/truly_transparent_text_with_pil.html

--Ned.

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

From jjhuff at mspin.net  Thu Jan 31 21:11:43 2008
From: jjhuff at mspin.net (Justin Huff)
Date: Thu, 31 Jan 2008 12:11:43 -0800
Subject: [Image-SIG] [patch] JPEG Chroma Subsampling
Message-ID: <47A22B7F.6080709@mspin.net>

Attached is a patch (against PIL 1.1.6) to provide control over the
chroma subsampling done by the JPEG encoder.  This is often useful for
reducing compression artifacts around edges of clipart and text.

Comments appreciated!

--Justin
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: PIL-jpeg-subsampling.patch
Url: http://mail.python.org/pipermail/image-sig/attachments/20080131/d3b53da8/attachment.txt 

From seed95 at gmail.com  Wed Jan 30 09:51:42 2008
From: seed95 at gmail.com (Seed)
Date: Wed, 30 Jan 2008 09:51:42 +0100
Subject: [Image-SIG]  Make the difference between a picture and a photo
In-Reply-To: <52e42efe0801291656i3b8c149cn6b4e2190f0c40527@mail.gmail.com>
References: <52e42efe0801291656i3b8c149cn6b4e2190f0c40527@mail.gmail.com>
Message-ID: <52e42efe0801300051saa0721bre672ffac18c6541d@mail.gmail.com>

Hi,

Is there a way with or  without PIL python librairy to detect if an image
looks like a
picture more than a photography and the opposite?

Any  help will be appreciated.

Thanks in advance.

-- 
Sidney Amani
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20080130/c912362a/attachment.htm