From Zachary_Roadhouse@brown.edu Sat Oct 5 16:32:57 1996 From: Zachary_Roadhouse@brown.edu (Zachary_Roadhouse@brown.edu) Date: Sat, 05 Oct 1996 11:32:57 -0400 (EDT) Subject: [PYTHON IMAGE-SIG] When will PIL v0.2 be released? Message-ID: I heard it said that PIL v0.2 would be released sometime during september. What is its status? - Zack E-MAIL: Zachary_Roadhouse@brown.edu WEB: http://www.lems.brown.edu/~zack Box 220, Brown University, Providence, RI 02912 Phone: (401) 863 - 5435 ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org ================= From fredrik_lundh@ivab.se Sat Oct 5 16:46:41 1996 From: fredrik_lundh@ivab.se (Fredrik Lundh) Date: Sat, 5 Oct 1996 17:46:41 +0200 Subject: [PYTHON IMAGE-SIG] When will PIL v0.2 be released? In-Reply-To: Message-ID: <9610051546.AA02025@arnold.image.ivab.se> > I heard it said that PIL v0.2 would be released sometime > during september. What is its status? Make that early October :-) As it turned out, I'll be extremely busy with other things until the end of next week. My plans were to release a "0.2 alpha" yesterday (the README and change log is already available at www.python.org), but the final testing exposed a critical bug that I haven't fixed yet. Will make another attempt tomorrow. Regards /F ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org ================= From Zachary_Roadhouse@brown.edu Sun Oct 6 03:40:57 1996 From: Zachary_Roadhouse@brown.edu (Zachary_Roadhouse@brown.edu) Date: Sat, 05 Oct 1996 22:40:57 -0400 (EDT) Subject: [PYTHON IMAGE-SIG] Changes in PIL 0.2a1 Message-ID: I couldn't read the readme file (server said it wasn't there), but I liked the rest of the features that have been added. Some of them, specifically the pixel manipulation routines and postscript fixes are particularly usefull for the Raphael project that our group at LEMS was working on. Now if I can just get Dale to release his Image Magick module... - Zack E-MAIL: Zachary_Roadhouse@brown.edu WEB: http://www.lems.brown.edu/~zack Box 220, Brown University, Providence, RI 02912 Phone: (401) 863 - 5435 ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org ================= From fredrik_lundh@ivab.se Sun Oct 6 15:51:10 1996 From: fredrik_lundh@ivab.se (Fredrik Lundh) Date: Sun, 6 Oct 1996 16:51:10 +0200 Subject: [PYTHON IMAGE-SIG] Changes in PIL 0.2a1 In-Reply-To: Message-ID: <9610061451.AA09623@arnold.image.ivab.se> Zack wrote: > I couldn't read the readme file (server said it wasn't there), but I > liked the rest of the features that have been added. Okay, just fixed the readme, and uploaded the first "snapshot" alpha. Download and try it out at your own risk ;-) Expect more features to be included (or simple enabled) within the next few weeks. Cheers /F ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org ================= From Dirk.Engelmann@iwr.uni-heidelberg.de Thu Oct 17 15:49:41 1996 From: Dirk.Engelmann@iwr.uni-heidelberg.de (Dirk Engelmann) Date: Thu, 17 Oct 1996 15:49:41 +0100 Subject: [PYTHON IMAGE-SIG] convolution Message-ID: <9610171549.ZM19988@chagall.iwr.uni-heidelberg.de> Hello, I would like to know, if there is an efficient method for python available to calculate a convolution (of a 2D image and a a windowing mask). This would be quite usefull for filter design. Also I would like to know, if there are operators like fast fourier transforms (laplace transforms, etc.). Thanks for any hints! regards, Dirk Engelmann ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org ================= From fredrik_lundh@ivab.se Fri Oct 18 09:43:03 1996 From: fredrik_lundh@ivab.se (Fredrik Lundh) Date: Fri, 18 Oct 1996 10:43:03 +0200 Subject: [PYTHON IMAGE-SIG] convolution In-Reply-To: <9610171549.ZM19988@chagall.iwr.uni-heidelberg.de> (Dirk.Engelmann@iwr.uni-heidelberg.de) Message-ID: <9610180843.AA03028@arnold.image.ivab.se> (intentionally cross-posted to matrix-sig) > I would like to know, if there is an efficient method for python > available to calculate a convolution (of a 2D image and a a > windowing mask). The PIL itself supports only 3x3 and 5x5 convolution filters, mainly intended for "image enhancement" purposes. In 0.1, there's only a few hardwired filters available, you cannot provide you own kernels (this will change in 0.2). For fancier stuff, convert the image into a matrix and use numerical python to process it (the most recent version of NumPy includes FFT support). A crude interface between PIL and NumPy (using tostring) interface has been posted to this list earlier; PIL 0.2 will (hopefully) include slicker interfaces. As for 2D convolutions, ask on the matrix-sig list. It's full of number crunching wizards (right? ;-) Regards /F ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org ================= From ve6ahm@ve6ahm-1.ampr.ab.ca Sat Oct 26 19:23:12 1996 From: ve6ahm@ve6ahm-1.ampr.ab.ca (Les Davies) Date: Sat, 26 Oct 1996 11:23:12 -0700 (MST) Subject: [PYTHON IMAGE-SIG] PyImagingCore - TkImaging.c Message-ID: <199610261823.LAA03798@ve6ahm-1.ampr.ab.ca> I have been trying to use the tkinter to display images however the correct method to install TkImaging.c to get PyImagingcore recognized as a valid Imaging format in Tk seems to be eluding me. Any help would be appreciated. My last attempt was to compile it with _tkinter in the modules directory of Python1.4b3. Tkappinit was modified and included with the above compile. Pilview and my own program both fail with PyImagingCore not found. I can provide dumps if required of the error messages if additional info required. This is with the newsest Imaging modules. Les ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org ================= From ve6ahm@ve6ahm-1.ampr.ab.ca Sun Oct 27 13:19:01 1996 From: ve6ahm@ve6ahm-1.ampr.ab.ca (Les Davies) Date: Sun, 27 Oct 1996 06:19:01 -0700 (MST) Subject: [PYTHON IMAGE-SIG] PyImagingCore - TkImaging.c In-Reply-To: <199610261823.LAA03798@ve6ahm-1.ampr.ab.ca> from "Les Davies " at Oct 26, 96 11:23:12 am Message-ID: <199610271319.GAA12119@ve6ahm-1.ampr.ab.ca> > > I have been trying to use the tkinter to display images however the > correct method to install TkImaging.c to get PyImagingcore recognized > as a valid Imaging format in Tk seems to be eluding me. Any help would be > appreciated. My last attempt was to compile it with _tkinter in the > modules directory of Python1.4b3. Tkappinit was modified and included with > the above compile. Pilview and my own program both fail with > PyImagingCore not found. > I can provide dumps if required of the error messages if additional info > required. This is with the newsest Imaging modules. > Les > Well, it seems I have answered my own question at about three in the morning here. For Linux2.0.6 with Caldera Preview 2 on top of Redhat 3.0.3 the Setup file would require: # *** ALWAYS enable this line: TKPATH=:$(DESTLIB)/tkinter # *** Enable *one* of the following lines: # For Solaris: #_tkinter _tkinter.c -I/usr/openwin/include -I/usr/local/include -L/usr/openwin/lib -L/usr/local/lib -ltk4.1 -ltcl7.5 -lX11 # For generic system (may have to add -I/-L options to find X11): _tkinter _tkinter.c tkappinit.c tkImaging.c -I../Extensions/Imaging/libImaging -I/usr/include/tcl -DWITH_APPINIT -L/usr/X11R6/lib -L/usr/lib -ltk4.0 -ltcl7.4 -lX11 This assumes that tkImaging.c has been copied to the Modules directory. Also it should be complied as a module which is not the default in Modules Setup file as you get a lot more debugging information if your includes are wrong. Some of the problems to avoid are: By default the makes required by Makefile.pre.in are not done for shared libraries. If you change Setup do not forget to do a make Makefiles to create a new Makefile. Hopefully this will help some one else. Les ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org ================= From fredrik_lundh@ivab.se Fri Oct 18 09:43:03 1996 From: fredrik_lundh@ivab.se (Fredrik Lundh) Date: Fri, 18 Oct 1996 10:43:03 +0200 Subject: [PYTHON IMAGE-SIG] convolution In-Reply-To: <9610171549.ZM19988@chagall.iwr.uni-heidelberg.de> (Dirk.Engelmann@iwr.uni-heidelberg.de) Message-ID: <9610180843.AA03028@arnold.image.ivab.se> (intentionally cross-posted to matrix-sig) > I would like to know, if there is an efficient method for python > available to calculate a convolution (of a 2D image and a a > windowing mask). The PIL itself supports only 3x3 and 5x5 convolution filters, mainly intended for "image enhancement" purposes. In 0.1, there's only a few hardwired filters available, you cannot provide you own kernels (this will change in 0.2). For fancier stuff, convert the image into a matrix and use numerical python to process it (the most recent version of NumPy includes FFT support). A crude interface between PIL and NumPy (using tostring) interface has been posted to this list earlier; PIL 0.2 will (hopefully) include slicker interfaces. As for 2D convolutions, ask on the matrix-sig list. It's full of number crunching wizards (right? ;-) Regards /F ================= IMAGE-SIG - SIG on Image Processing with Python send messages to: image-sig@python.org administrivia to: image-sig-request@python.org =================