[Image-SIG] Resampling with antialiasing while using only 8 greyscales
Good evening, F irstly, thanks for creating and maintaining such a fantastic library, the PIL is awesome! I've got an interesting problem and I'd really appreciate if someone could point me in the right direction. I'm a reasonable developer but I've become out of my depth in image science :) I've got greyscale images which I have dithered (Floyd-Steinberg) so that they contain exactly 8 shades. The pixel values present in the images are 0, 36, 73, 109, 146, 182, 219, 255 - so 8 levels with pure-white and pure-black being represented. Naturally, most of the 256 possible pixel values in the histogram are empty. I want to reduce the size of this image by 3x in both directions, ie (300, 300) => (100, 100) *with* anti-aliasing but *without* creating pixels outside of those 8 shades. Of course, the Lanczos implementation in PIL doesn't know to only use values 0, 36, 73 etc, so it uses the full 256 range, thereby undoing my dithering work. Can anyone suggest a good approach to this? Would it be possible to modify the resampling algorithm to work in 8 shades rather than 256? Or is there a way I could put the pixel values back into the 8 "buckets" afterwards without wrecking the colour accuracy? I have tried the latter and gradients become steps, as you'd expect. It doesn't seem right to apply dithering again somehow. I have considered re-ordering my process from: *High-res image -> Dithering -> Resampling* to *High-res image -> Resampling -> Dithering* but I then think the dithering process wreaks havoc on the beautifully antialiased edges in the image. Any thoughts would be very gratefully received. Kind regards, Paul ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] [SPAM] - Re: fromarray rotates image - Email found in subject
The second image is a 32 bit tif, so you need to use something like ImageJ or
Fiji. Microsoft image viewer won't work. I think I shouldn't have to fiddle
with the row/column order, but I guess it needs to stay the way it is. Thanks
for the help.
Paul
-Original Message-
From: Chris Barker [mailto:chris.bar...@noaa.gov]
Sent: Tuesday, January 31, 2012 1:18 PM
To: Cousoulis, Paul; image-sig@python.org
Subject: [SPAM] - Re: [Image-SIG] fromarray rotates image - Email found in
subject
On Tue, Jan 31, 2012 at 8:38 AM, Cousoulis, Paul
wrote:
> I'm sorry but I still think there is a bug.
I still don't think so: explanation below.
By the way, there is another problem with your example -- I get an all-black
second image. I think you need to specify the image mode when you create the
new image:
newimage = Image.fromarray(npArray, mode=image1.mode)
though that still makes a mess of it! -- more debugging to be done here -- see
below
> In [4]: print image1.size
> (516, 356)
>
> In [5]: npArray = np.array(image1.getdata())
>
> In [6]: print npArray.shape
> (183696L,)
OK so far
> In [7]: npArray.shape = image1.size
> In [8]: print npArray.shape
> (516L, 356L)
here I would swap -- as numpy naturally stores data the other way:
and now it works:
In [8]: run pil-numpy-test.py
input image size: (516, 356) numpy image shape before: (183696,) numpy image
shape after: (356, 516) new image size (516, 356)
(though the colors are still screwed up -- I don't know what's up with that...)
I can see how you'd expect it to work the way you had it, but I think the
problem is that you are mixing two ways to push raw data to/froim numpy arrays:
npArray = np.array(image1.getdata())
is using PIL's getdata() to put the raw data in a string, then turning that
into a numpy array (oh, and that may be the source of teh data mess up
too...np.array is expecting a sequence of numbers or something, not raw data --
you want:
OOPS, actually, getdata returns something else altogether:
"""
getdata
im.getdata() => sequence
Returns the contents of an image as a sequence object containing pixel values.
The sequence object is flattened, so that values for line one follow directly
after the values of line zero, and so on.
Note that the sequence object returned by this method is an internal PIL data
type, which only supports certain sequence operations, including iteration and
basic sequence access. To convert it to an ordinary sequence (e.g. for
printing), use list(im.getdata()).
"""
so it should be:
npArray = np.fromstring(image1.tostring(), dtype=np.uint8)
npArray.shape = (image1.size[1], image1.size[0] )
and that now works -- correct size, and correct final image.
However:
newimage = Image.fromarray(npArray)
is using the numpy "array protocol", which is a bit different than
fromstring/tostring -- it carries shape information -- hence the need to
specify the shape of the numpy array as I did.
you can use that protocol both ways:
npArray = np.asarray(image1)
which then preserved size info.
Here's my new version:
from PIL import Image
import numpy as np
image1 = Image.open("LineGraph.tif")
print image1
print "input image size:", image1.size
npArray = np.asarray(image1)
print "numpy image shape after:", npArray.shape
newimage = Image.fromarray(npArray, mode=image1.mode)
print newimage
print "new image size", newimage.size
newimage.save("LineGraph2.tif")
NOTE: if you do fromstring/tostring (or tobuffer) consistently, then it doesn't
matter what shape you make the numpy array:
image1 = Image.open("LineGraph.tif")
print image1
print "input image size:", image1.size
npArray = np.fromstring(image1.tostring(), dtype=np.uint8)
print "numpy image shape:", npArray.shape
newimage = Image.fromstring(image1.mode, image1.size, npArray.tostring())
print newimage
print "new image size", newimage.size
newimage.save("LineGraph2.tif")
But that's less efficient, and messier.
NOTE: it might have been nicer if the array protocol were used such that the
array created was fortran-order, and thus (w,h) in shape, but so it goes.
HTH,
-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.bar...@noaa.gov
==
This electronic message transmission and any attachments are
confidential and/or proprietary and may constitute legally privileged
information of
Meso Scale Diagnostics, LLC. The information is intended for solely
the use of recipient (recipient). If you are not
the intended recipient, yo
[Image-SIG] fromarray rotates image
The fromarray method is rotating the shape of arrays when converting from numpy arrays. In [56]: npArray.shape Out[56]: (650, 670) In [57]: newimage = Image.fromarray(npArray) In [58]: newimage.size Out[58]: (670, 650) In [59]: Image.VERSION Out[59]: '1.1.7' Thanks Paul == This electronic message transmission and any attachments are confidential and/or proprietary and may constitute legally privileged information of Meso Scale Diagnostics, LLC. The information is intended for solely the use of image-sig@python.org (image-sig@python.org). If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution or the taking of any action in reliance of this information is strictly prohibited. You are not authorized to retain it in any form nor to re-transmit it, and you should destroy this email immediately. If you have received this electronic transmission in error, please notify us by telephone (240-631-2522) or by electronic mail to the sender of this email, Cousoulis, Paul (pcousou...@meso-scale.com), immediately. = ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] PIL does not support iTXt PNG chunks [patch]
(Forwarded from: http://bugs.python.org/issue13514 per Ezio Melotti)
The Python Imaging Library does not support handling of UTF-8 'iTXt'
key:value chunks in PNG files:
http://www.w3.org/TR/PNG/#11iTXt
Such support is necessary for successful extraction of key:value pairs
of UTF-8 encoded data, stored in an PNG 'iTXt' comment chunk.
The following example file (from British GCHQ) demonstrates such a
record in a PNG file. Based on this evidence, it is highly likely
that GCHQ hide all of their important secrets using this kind of
steganography, and likely necessary that spies from the rest of the
world are requiring similar access to GCHQ secrets. Inclusion of a
working chunk_iTXt() PIL/PNG support function will enable more
harmonious and effective eavesdropping. Example image:
http://www.canyoucrackit.co.uk/images/cyber.png
(The attached .py file is not a directly apply-able patch, but
contains a working implementation for chunk_iTXt() and a demonstrative
test function for inserting it).
-Paul
#!/usr/bin/env python
# Paul Sladen, 2011-12-01
from PIL import PngImagePlugin, Image, ImageFile
#Image.DEBUG = True
# Meh, why doesn't PIL support this already?
def chunk_iTXt(self, pos, len):
s = ImageFile._safe_read(self.fp, len)
# http://www.w3.org/TR/PNG/#11iTXt
k, structure = s.split('\0', 1)
compression_flag, compression_method = map(ord, structure[:2])
language_tag, translated, t = structure[2:].split('\0', 2)
if compression_flag == 0:
v = t.decode('utf-8')
elif compression_flag == 1 and compression_method == 0:
import zlib
v = zlib.decompress(t).decode('utf-8')
else:
raise SyntaxError("Unknown compression request %d in iTXt chunk" %
compression_method)
self.im_info[k] = {'translated_keyword': translated.decode('utf-8'),
'language_tag': language_tag,
'text': v}
self.im_text[k] = v
return s
PngImagePlugin.PngStream.chunk_iTXt = chunk_iTXt
def test(f):
i = PngImagePlugin.PngImageFile(f)
print i.info, i.text
steno = i.text['Comment'].decode('base64')
if __name__ == '__main__':
test('cyber.png')
___
Image-SIG maillist - Image-SIG@python.org
http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Mac support for PIL
Thanks, and sorry, what I meant to say is that PIP is really easy to install on windows but a bit of a trial on Mac. There are loads of blog entries around, which perhaps reflects the trouble that people are having. Many of these entries are highly version specific and have a less than unity probability of success (I tried three before I managed to get things working). Judging by the comment on most blogs other people are in the same boat as I. FWIW this is what eventually worked for me: Based on: http://rezmuh.sixceedinc.com/2009/04/setting-up-pil-with-libjpeg-on-mac-os-x-leopard.html I got and installed MacPorts-1.8.2-10.6-SnowLeopard.dmg from: http://distfiles.macports.org/MacPorts/ Then in Imaging-1.1.7: In setup.py changed the JPEG_ROOT to: JPEG_ROOT = "/opt/local/lib/libjpeg.dylib" Then: sudo python setup.py build --force sudo python setup.py install I would not guarantee this will work for others! Thanks for the help and thanks for such a great product as PIL Regards, Paul. FWIW I s On 7 May 2010, at 16:05, Yury V. Zaytsev wrote: > Hi! > > On Wed, 2010-05-05 at 21:42 +0100, Paul Ross wrote: >> >> I know that PIL is Windows-centric but is there any chance that it will >> support Snow Leopard on Apple Mac some time soon? > > Windows-centric?! > >> I have googled for the answer for a couple of hours now and tried many >> ad-hoc things (deleting PIL, installing libjpeg etc. etc.) but I >> always just get "IOError: decoder jpeg not available" whenever I try >> and manipulate JPEG images > > Basically you have to build and install libjpeg and other decoders from > source or use a package management system, such as Fink or Macports, > then rebuild PIL against them. > > If you want more help, ask specific questions. > > -- > Sincerely yours, > Yury V. Zaytsev > ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Mac support for PIL
Hi there and I'm sorry to bother you but... I know that PIL is Windows-centric but is there any chance that it will support Snow Leopard on Apple Mac some time soon? I have been trying to build PIL from source on Snow Leopard without success as it appears to only support PNG and not other image formats such as JPEG. I have googled for the answer for a couple of hours now and tried many ad-hoc things (deleting PIL, installing libjpeg etc. etc.) but I always just get "IOError: decoder jpeg not available" whenever I try and manipulate JPEG images Many thanks if you can help me hack PIL to build on a Mac! Paul. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] PIL 1.1.7 binary for python 2.4
Hi Fredrik, I'm using PIL to do some quick image manipulation and I'm very happy with it. I would like to use the chroma subsampling that is present in PIL 1.1.7, but I'm restricted to use Python 2.4... Could you point me out where I can find a Windows binary of PIL 1.1.7 for Python 2.4? Thanks! Best regards, Paul ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] ImageFont family mojibake
2009/9/16 Donn : > Thanks to Stani for letting me know I did reach the list. Seems I disabled my > receiving of email from the list a while back and I forgot about it. > > I will get mail from now on, but am not sure if anyone weighed-in on the OP. > Is there any feedback? If so, is there a link to the archives somewhere? > > Thanks, sorry for the noise. > \d I've seen 3 mails from you and no replies. Looks like nobody's been able to help. Paul ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] ImageFont family mojibake
2009/9/16 Donn : > Another ping. Is anyone seeing the OP? > \d I see it. Just can't help... ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] ImageTk.PhotoImage: DLL load failed: The specified module could not be found.
Hello, I have an problem with ImageTk.PhotoImage. When I try to use it with for example with the following script: import ImageTk import Image import tkFileDialog imgFile = tkFileDialog.askopenfilename() if imgFile: image = Image.open(imgFile) chart = ImageTk.PhotoImage(image) I got the following error: Traceback (most recent call last): File "C:\Python\PythonScript\Test\Script1.py", line 9, in chart = ImageTk.PhotoImage(image) File "C:\Python25\Lib\site-packages\PIL\ImageTk.py", line 116, in __init__ self.paste(image) File "C:\Python25\Lib\site-packages\PIL\ImageTk.py", line 181, in paste import _imagingtk ImportError: DLL load failed: The specified module could not be found. The file _imagingtk is present in the path C:\Python25\Lib\site-packages\PIL\. I don't understand what is wrong. I am using ActivePython 2.5.4.4 (ActiveState Software Inc.)base on Python 2.5.4 (r254:67916, Apr 27 2009, 15:41:14). Thanks for your help, Jean-Paul ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] ImageTk.PhotoImage: DLL load failed: The specified module could not be found.
Hello, I have an problem with ImageTk.PhotoImage. When I try to use it with for example with the following script: import ImageTk import Image import tkFileDialog imgFile = tkFileDialog.askopenfilename() if imgFile: image = Image.open(imgFile) chart = ImageTk.PhotoImage(image) I got the following error: Traceback (most recent call last): File "C:\Python\PythonScript\Test\Script1.py", line 9, in chart = ImageTk.PhotoImage(image) File "C:\Python25\Lib\site-packages\PIL\ImageTk.py", line 116, in __init__ self.paste(image) File "C:\Python25\Lib\site-packages\PIL\ImageTk.py", line 181, in paste import _imagingtk ImportError: DLL load failed: The specified module could not be found. The file _imagingtk is present in the path C:\Python25\Lib\site-packages\PIL\. I don't understand what is wrong. I am using ActivePython 2.5.4.4 (ActiveState Software Inc.)base on Python 2.5.4 (r254:67916, Apr 27 2009, 15:41:14). Thanks for your help, Jean-Paul ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Resizing Photos
Hi, Scratch that - I found the ANTIALIAS mode and the thumbnails are now much better quality. Thanks for a great lib! Paul On Tue, Apr 21, 2009 at 9:11 AM, Paul Johnston wrote: > Hi, > > I'm currently using PIL to resize photos down to a thumbnail. Something > I've noticed though, is that whatever software Facebook use too do the same, > produces better-quality thumbnails. > > Are there any tweaks I can apply to PIL to get better resizing? Or any idea > what software FB may be using? > > Paul > ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Resizing Photos
Hi, I'm currently using PIL to resize photos down to a thumbnail. Something I've noticed though, is that whatever software Facebook use too do the same, produces better-quality thumbnails. Are there any tweaks I can apply to PIL to get better resizing? Or any idea what software FB may be using? Paul ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] PIL 1.1.6 and convert cmyk jpegs to rgb and writing png files
Just in case this CMYK patch does make it into the PIL svn repo...
there's a small bug: before calling invert, self.im should be checked.
This is necessary because the ImageFile.Parser.feed method tries an
open() to see if the image is complete.
Attached is an updated patch.
Index: PIL/JpegImagePlugin.py
===
--- PIL/JpegImagePlugin.py (revision 460)
+++ PIL/JpegImagePlugin.py (working copy)
@@ -1,6 +1,6 @@
#
# The Python Imaging Library.
-# $Id: JpegImagePlugin.py 2763 2006-06-22 21:43:28Z fredrik $
+# $Id: JpegImagePlugin.py 2199 2004-12-18 08:49:05Z fredrik $
#
# JPEG (JFIF) file handling
#
@@ -32,7 +32,7 @@
__version__ = "0.5"
import array, string
-import Image, ImageFile
+import Image, ImageFile, ImageChops
def i16(c,o=0):
return ord(c[o+1]) + (ord(c[o])<<8)
@@ -270,8 +270,11 @@
handler(self, i)
if i == 0xFFDA: # start of scan
rawmode = self.mode
-if self.mode == "CMYK":
-rawmode = "CMYK;I"
+# patch by Kevin Cazabon to comment this out - nobody should be using Photoshop 2.5 any more (and it breaks newer versions)
+# CMYK images are still inverted, we'll fix that just before returning.
+#if self.mode == "CMYK" and self.info.has_key("adobe"):
+#rawmode = "CMYK;I" # Photoshop 2.5 is broken!
+
self.tile = [("jpeg", (0,0) + self.size, 0, (rawmode, ""))]
# self.__offset = self.fp.tell()
break
@@ -282,6 +285,10 @@
else:
raise SyntaxError("no marker found")
+# patch by Kevin Cazabon to re-invert CMYK JPEG files
+if self.im and self.mode == "CMYK":
+self.im = ImageChops.invert(self).im
+
def draft(self, mode, size):
if len(self.tile) != 1:
@@ -378,7 +385,7 @@
"RGB": "RGB",
"RGBA": "RGB",
"RGBX": "RGB",
-"CMYK": "CMYK;I",
+"CMYK": "CMYK",
"YCbCr": "YCbCr",
}
@@ -406,6 +413,10 @@
dpi[0], dpi[1]
)
+if im.mode == "CMYK":
+# invert it so it's handled correctly in Photoshop/etc. - Kevin Cazabon.
+im = ImageChops.invert(im)
+
ImageFile._save(im, fp, [("jpeg", (0,0)+im.size, 0, rawmode)])
def _save_cjpeg(im, fp, filename):
___
Image-SIG maillist - Image-SIG@python.org
http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Font colour.
On Thu, 26 Apr 2007, Amos Newcombe wrote: > You find the complementary color by subtracting each color band from 255 > (assuming 8-bit color). The problem with this is that if your original color > is medium gray, the complementary color will be very close to it and you > will get no legibility. So you have to test for this case and go to some > other color, perhaps black or white. Slightly better than complementing each color band, set it as far off from the background as you can; [00..7F] -> FF, [80..FF] -> 00. /Paul > Amos > > On 4/26/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: >> >> Christopher , >> Thank you for your reply. >> >> Is there a way how to find out a complementary colour for an area where I >> will write the text, so >> that the text will be seen clearly?Is there a routine in PIL or in Python >> somewhere? >> Thank you for help >> Lad. >> > -- ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] PIL on a Mac
I sent a how-to using finkl as a sandbox to perform the installation. However, it was never sent through. Why? On 4/24/07, Bill Janssen <[EMAIL PROTECTED]> wrote: > It does, but it's an old version But not significantly different from the current version in most respects, certainly for a beginner. I've built and still use large complicated apps with lots of external Python libraries that run just fine on the system Python. > and somewhat broken in some subtle ways. Too subtle to be interesting. Bill ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig -- http://covertheuninsured.org/ http://www.nrdcaction.org/gwtakeaction http://www.nrdcactionfund.org/tellafriend.asp http://www.savetheinternet.com ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] PIL on a Mac
I sent a how-to using finkl as a sandbox to perform the installation. However, it was never sent through. Why? On 4/24/07, Bill Janssen <[EMAIL PROTECTED]> wrote: > It does, but it's an old version But not significantly different from the current version in most respects, certainly for a beginner. I've built and still use large complicated apps with lots of external Python libraries that run just fine on the system Python. > and somewhat broken in some subtle ways. Too subtle to be interesting. Bill ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig -- http://covertheuninsured.org/ http://www.nrdcaction.org/gwtakeaction http://www.nrdcactionfund.org/tellafriend.asp http://www.savetheinternet.com ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] PIL on a Mac
Hi, You should checkout http://www.finkproject.org/ which creates a 'sandbox' for gnu/unix/non-mac utilities. It comes with a GUI which lets you select packages from their repository (similar to a yum/apt-get repository for several linux distributions out there). It also comes with command-line tools, if you prefer. Once you've installed fink and python, you can directly call the python binary installed in the fink default bin (eg, /sw/bin/) when using the setup tools (eg, setup.py) which comes with PIL. Using this fink 'python' will install the PIL package inside python's corresponding site-packages folder. In order to set the fink environment automatically every time you use the terminal, you'll need to modify your .bash_profile. Add the following line "test -r /sw/bin/init.sh && . /sw/bin/init.sh". You'll of course need to adapt the fink installation path as you see fit. Good luck, Paul On 4/13/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: Hello Support, I'm sitting here looking at the ReadMe for the PIL, and I realized something a few moments back: this is more complicated than I thought it would be. If it isn't too difficult, would you be able to tell me how I could put the PIL in my computer so I can just play around with Python and images. I am currently in a introductory course to computer science and we are using Python. I am running it on a MacBook Pro right now. I guess I had to download a MacPython in order to make Python work on my system. I'm not too sure, but it does seem to work. The problem that I'm having now seems to be stemming from his "Imaging-1.1.6" folder that I just downloaded. What should I do with it? I was looking at the ReadMe (that's where I got this e-mail) and it says that I'll need to download more things (correct?). My professor did say something about Fink. Now, I don't want you guys to go totally out of your way because there is a computer lab on campus that I could do all this on, but I would like it to be on my computer, too, just so I can finish the homework up and play around with image transformations. Thank you. Adam Crouse ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig -- http://www.nrdcaction.org/gwtakeaction http://www.nrdcactionfund.org/tellafriend.asp http://www.savetheinternet.com ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Crack Code?
Hi, Whatever happened to the PIL module CrackCode?? PIL plus is no longer available either. Thanks, Paul -- http://www.nrdcactionfund.org/tellafriend.asp ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] CrackCode
Hi, Whatever happened to the CrackCode PIL module? Thanks, Paul -- http://www.nrdcactionfund.org/tellafriend.asp ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] xpm files not supported
Hi all, There was an old post regarding xpm compatibility. I was wondering whether the latest development version has incorporated this users suggestions? http://mail.python.org/pipermail/image-sig/2006-October/004159.html I'm currenly using 1.1.6 and the suggested fixes haven't been incorporated yet. Are there any plans on doing so? Thanks, Paul -- http://www.nrdcactionfund.org/tellafriend.asp ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Newb help... can't install the PIL extension libraries
Hi, I'm very new to linux (2 days!) but have been using Python on windows for a few weeks now. I would like to install PIL on my linux (ubuntu) machine but am having all sorts of trouble following the readme, I get a ton of errors when trying to run the build.py. Mostly cannot find directory errors, but there are others thrown in for good measure. This is my first attempt at importing a library within linux. I unzipped the tar file into my user space... /home/paul/Python/ExtensionLibraries which I created myself. This is where I'm going wrong I presume. The readme said that I could create my own directory, so I'm confused. Should I unzip it & build it somewhere else? Any help greatly appreciated, many thanks. PMF ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] PIL Programming Questions
I just finished a PIL program and I feel like there may have been more
efficient ways to do things I did "by hand" with the raw data. Can
someone enlighten me?
1. I wanted to take 24 bit data, count the colours and make an 8-bit BMP
if there are fewer than 256 colours. My brute force solution required me
to build the palette by hand from the data and rewrite the data one
pixel at a time. (a simple '.convert("P")' seems to use an arbitrary
palette or something???)
2. I wanted to pad a bitmap by adding some space to the right and
bottom. I did this by creating a new one and pasting the old one into
it.
3. I wanted to convert every occurrence of one colour to another. I
walked each pixel.
I also could not figure out how to get to a 4-bit BMP. I inferred from
the documentation that this just isn't possible with out of box PIL.
Overall: PIL really rocks.
Paul Prescod
___
Image-SIG maillist - Image-SIG@python.org
http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Making the "outside" of an image transparent
On 5/10/06, Fredrik Lundh <[EMAIL PROTECTED]> wrote: > Paul Moore wrote: > > > So I can get transparency set, but I still don't know a good flood > > fill algorithm. > > eric raymond posted a nice implementation last year, which will go > into 1.1.6: > > http://article.gmane.org/gmane.comp.python.image/1753 > > (note that his code doesn't work on PIL Image objects; mapping this > to corresponding PIL methods is left as an exercise etc etc) That's the one I found, and based my code on. The problem I had was that it floods everything until it hits a specific colour - whereas I am looking to flood evertything of one colour, until I hit a different colour. It's almost the same, but something in the edge conditions went wrong, I think. Actually, things were made worse by the fact that I was creating images for a web page, and I hadn't realised that Internet Explorer only seems to respect transparency in GIF files. (I was avoiding GIF, as PIL doesn't compress them, for copyright reasons, as I understand it). I'll go back and look again and see if I can get that algorithm working. Thanks, Paul. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Making the "outside" of an image transparent
On 5/5/06, Paul Moore <[EMAIL PROTECTED]> wrote: > I searched for a flood fill algorithm that would do what I wanted, and > couldn't find anything that I could use. I did find one algorithm, > which I tried to modify to do what I wanted, but I've clearly > misunderstood something vital, as the resulting image looked identical > to the original :-( I managed to get the transparency side of things working - the big problem was that I didn't realize GIF transparency worked differently from other image formats, so I was doing the wrong thing there. So I can get transparency set, but I still don't know a good flood fill algorithm. Thanks, Paul. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Making the "outside" of an image transparent
I have a rectangular image, which is a logo on a white background. I
would like to modify the image to make the white background
transparent (in effect, giving an irregularly shaped image). I've
searched the archives, and found some details on how to make parts of
an image transparent, but they aren't quite what I need. The problem I
have is that parts of the interior of the image are also white, and
those parts I don't want to make transparent.
I searched for a flood fill algorithm that would do what I wanted, and
couldn't find anything that I could use. I did find one algorithm,
which I tried to modify to do what I wanted, but I've clearly
misunderstood something vital, as the resulting image looked identical
to the original :-(
Can anyone help me? I attach my code in case it's of use, but to be
honest, I'm not sure it's even close...
Thanks,
Paul.
import Image
def crop_background(image, colour = None):
w, h = image.size
print w*h
mask = Image.new("L", image.size)
if colour is None:
colour = image.getpixel((0, 0))
pt = 0, 0
background = set()
seen = set()
candidates = set((pt,))
while candidates:
pt = candidates.pop()
seen.add(pt)
if image.getpixel(pt) == colour:
background.add(pt)
i, j = pt
for new_i, new_j in (i+1, j), (i, j+1), (i-1, j), (i, j-1):
if 0 <= new_i < w and 0 <= new_j < h and (new_i, new_j) not in seen:
candidates.add((new_i, new_j))
pix = mask.load()
for i in range(w):
for j in range(h):
if (i, j) in background:
pix[i, j] = 0
else:
pix[i, j] = 1
image.putalpha(mask)
print len(background)
return image
___
Image-SIG maillist - Image-SIG@python.org
http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Python 2.5 binaries?
On 4/13/06, Paul Moore <[EMAIL PROTECTED]> wrote: > Does anyone have a binary installer for PIL with Python 2.5 on > Windows? I can try building it myself, but traditionally, I've had a > lot of trouble assembling all of the necessary imaging libraries, so I > thought I'd save myself some time if some kind sould had already done > this. I got this built. It's not particularly hard, it's just a matter of getting all the dependencies. I've had an offer of hosting the file, so if anyone else wants it, it'll be made available in due course. Paul. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Python 2.5 binaries?
Does anyone have a binary installer for PIL with Python 2.5 on Windows? I can try building it myself, but traditionally, I've had a lot of trouble assembling all of the necessary imaging libraries, so I thought I'd save myself some time if some kind sould had already done this. (Follow-up question - if no-one else has a binary, and I do build one, would anyone like to host it to save others the trouble?) Thanks, Paul ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Pyvox release 0.72 now available
Travis, > Thanks for releasing all of your work and for the generous license. My pleasure. > Do you know if and/or when your library will be interoperable with > NumPy? It's on my list of good things to do, along with 500 other items. It's been moving towards the top, since I'd like to be able to use Matplotlib with Pyvox. > The new array interface makes this quite easy. I really like the idea of having a common underlying protocol for all multi-dimensional arrays; I'm still looking at the details of this particular proposal. > Perhaps the PIL could also learn to export and import the array > interface to finally get copy-less data-transfer with NumPy arrays. Also a good idea. PIL supports many more image file formats that Pyvox ever will, so interoperability would be nice. > On a related note. It would be nice if we could figure out better > ways to work together on the basic object layers for image and > volume processing (in my mind that's what NumPy array's are most > useful for). Sounds good in theory, but I'm a bit unsure on the practical details. A good starting place would seem to be a common N-D array protocol. Do you have any other specific suggestions? Paul Hughett -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Pyvox release 0.72 now available
The source distribution kit for Pyvox version 0.72 is now available from http://www.med.upenn.edu/bbl/downloads/pyvox/ Pyvox is a Python extension module for processing volume images, particularly medical images; it also includes some examples and simple applications. Features added in release 0.72 include: The interface for constructing convolution kernels has been completely redesigned and now supports the dynamic modification of kernels. Internal types now have min and max attributes which contain the minimum and maximum possible finite positive values representable in that type. Pyvox is currently available as an alpha release under an open-source license and is written in ANSI C and designed to be easily portable to any Unix or Posix-compatible platform. Some programs also require the X Window System. Paul Hughett Paul Hughett, Ph.D.Research Associate Brain Behavior Laboratory 10th floor Gates Building Hospital of the University (215) 662-6095 (voice) of Pennsylvania(215) 662-7903 (fax) 3400 Spruce Street Philadelphia PA 19104 [EMAIL PROTECTED] A rose by any other name confuses the issue. -Patrick E. Raume -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Announcing another Python imaging package?
I have been developing an open-source Python extension module for volume image processing, particularly medical imaging. Would it be appropriate for me to announce new releases of that package on this list? Paul Hughett -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] apparent typo in pilfont doc
http://www.pythonware.com/library/pil/handbook/pilfont.htm I think "pdf" is supposed to say "pcf". If not, an explanation is needed. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] vector graphics Q
On 10/16/05, Dmitry (Dima) Strakovsky <[EMAIL PROTECTED]> wrote: > I haven't done any testing of bitmap to SVG myself, so this is not a > voice of experience :) > hope this helps: > > http://www.scale-a-vector.de/svg-test4-e.htm > http://autotrace.sourceforge.net/ Thanks for the linke - I'll have a look... Paul. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] vector graphics Q
On 10/16/05, Dmitry (Dima) Strakovsky <[EMAIL PROTECTED]> wrote: > Fredrick does your lib work with SVG? I believe that PIL is for bitmap images only, and as such, SVG isn't supported. If you look in the appendix "Image File Formats" of the PIL handbook (http://www.pythonware.com/library/pil/handbook/index.htm) you'll see that SVG isn't documented as being supported. > Is there a library out there that is particularly nice (has > resize,rotate,copy,paste type functions)for working with SVGs? I also have some interest in vector graphics processing. (Although it's fairly casual so far). One specific thing I would like to be able to do is to take a bitmap format of a line image (a scanned copy of a line drawing, to be specific) and convert it to vector format. I realise that this isn't a trivial task (edge detection, and so on) but are there any tools or libraries that will do this? I haven't found any support in PIL for this, but maybe I've missed something... Thanks, Paul. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Copying EXIF data from one file or PIL image to another?
On 6/5/05, Terry Carroll <[EMAIL PROTECTED]> wrote: > Can anyone tell me if it's fairly easy to copy a JPEG file's EXIF data to > another file. Even better would be if this is doable in PIL. > > My basic problem: > > I've written a small program that reads a JPEG image in PIL, and creates a > copy that is slightly modified (it overlays the image with a datestamp > taken from the EXIF data), and saves it as a new file. It all works > swimmingly well, except that the ne file has no EXIF data. I'd like to > copy the original image's EXIF data to the new image. You can *read* EXIF information via the _getexif() method of the image. Use the KEYS dictionary in the ExifTags module to map numeric keys to friendly names. For example, >>> d = dict((ExifTags.TAGS[k], v) for k, v in im._getexif().items()) >>> d['DateTimeOriginal'] '2002:08:07 17:07:59' Unfortunately, I don't know of a way to *write* EXIF information with PIL. Hope this helps (a bit) Paul. ___ Image-SIG maillist - Image-SIG@python.org http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Rendering Japanese fonts
On 4/20/05, Lars Yencken <[EMAIL PROTECTED]> wrote:
> I'm trying to render a Japanese font to an image, but I'm having some
> problems.
[...]
> desiredText = unicode('nothing 日本語 appears', 'utf8').encode('euc_jp')
I don't know if you are literally including Japanese characters in
your source code, as you appear to be doing here, but unless you have
an encoding declaration in your module, this won't work - by default,
Python source code only handles ASCII characters.
Try using Unicode escapes for the Japanese characters, or reading the
characters from a file using an appropriate codec stream reader. That
might help. (If it doesn't, I'm afraid I can't help you :-))
Paul.
___
Image-SIG maillist - Image-SIG@python.org
http://mail.python.org/mailman/listinfo/image-sig
Re: [Image-SIG] Compiling PIL for 2.4
On Mon, 13 Dec 2004 08:39:45 +, Voidspace <[EMAIL PROTECTED]> wrote: > I've just upgraded to Python 2.4. I've installed the free microsoft > optimising compiler and hacked distutils to use it - following the > instructions from http://www.vrplumber.com/programming/mstoolkit/ . It > works great and I've built a windows installer for PyCrypto 2.0. > > I'm attempting to compile PIL for python 2.4. The build instructions for > windows say : I posted some instructions to this list a couple of days ago for building PIL with mingw. You may like to try that option (as far as I can tell, it's not as messy as using the MS free compilers). Alternatively, Fredrik has now published a build of PIL 1.1.5b1 for Python 2.4, if that's of use. Regards, Paul. ___ Image-SIG maillist - [EMAIL PROTECTED] http://mail.python.org/mailman/listinfo/image-sig
[Image-SIG] Building PIL for Windows (Python 2.4)
I have successfully built PIL 1.1.5b1 for Python 2.4, using the free
Mingw implementation of gcc. Here are instructions, for anyone who is
interested. (Please note, there is one place I needed to modify the
PIL sources - I'm not sure if what I did was the best way, but maybe
Fredrik could review and apply something suitable in the core
sources...)
I have mingw 3.3.3 installed - you need a recent enough version to
support building binaries which link to msvcr71. You need to have
followed the instructions in the Python documentation ("Installing
Python Modules" section 6.2.2) on building libpython24.a.
You also need the libraries freetype, tiff, jpeg, zlib and libgw32c
from the GnuWin32 project (gnuwin32.sourceforge.net). Get the
"developer files" distributions, and unzip them into a suitable
location (I used "..\Libs" relative to my PIL build directory - the
instructions below will be based on this).
>From ..\Libs\lib, delete all of the .lib and .dll.a files - these link
to DLL versions of the libraries, and can get picked up by accident.
We want to link statically, to reduce dependencies.
I also have ActiveTCL (aspn.activestate.com) installed, for the TCL/TK
header files and link libraries. You need to copy tcl84.lib to
libtcl84.a, and tk84.lib to libtk84.a, to match mingw naming
conventions.
OK, that's the environment set up.
Now, we need to fix a problem in the PIL sources. Edit ImPlatform.h,
and comment out the definition of INT64. (This definition clashes with
others in some of the Windows header files, and I found that removing
it was the quickest fix to get everything to build...)
Now, in setup.py, set the roots:
FREETYPE_ROOT = libinclude("../Libs")
JPEG_ROOT = libinclude("../Libs")
TIFF_ROOT = libinclude("../Libs")
ZLIB_ROOT = libinclude("../Libs")
TCL_ROOT = libinclude("C:/Apps/Tcl") # Or wherever you installed ActiveTCL
Finally, you need to add a couple of dependencies to the _imagingft
extension (again, in setup.py):
exts.append(Extension(
"_imagingft", ["_imagingft.c"], libraries=["freetype",
"z", "gw32c"],
define_macros=defs
))
[The formatting is probably messed up here - it's the addition of "z"
and "gw32c" in the libraries argument that matters]
Now, you should just be able to build.
python setup.py build --compiler=mingw32 bdist_wininst
Paul.
PS I have a binary installer built. While I don't want to end up mass
mailing it to everyone, if anyone can offer space to host it, I'd be
happy to make it available. No guarantees beyond "it works for me",
unfortunately :-)
___
Image-SIG maillist - [EMAIL PROTECTED]
http://mail.python.org/mailman/listinfo/image-sig
