For review: IPL Coder

[seanburke1979]

Hello all.
I have a coder file that I would like reviewed for reading and writing IPLab files. (IPL) Please get a copy of the source at ftp://depot.aecom.yu.edu/pub/HaEF9Fzbf4C84DPMH304/ipl.c .
Corrections and suggestions are welcome.

Thanks!

Sean

[magick]

Thanks. Can you post a URL to a couple of IPL image files so we can test your code before we include it in the ImageMagick
distribution.

[seanburke1979]

8 bit single frame
8 bit multi frame
16 bit single frame
16 bit multi frame

[seanburke1979]

The links above are dead. If there is anyone still working on this coder, please send me an email and I will provide a fresh link for the files.

Best,

Sean

[seanburke1979]

What has been added:
-Support for writing U8, S16, S32, Float and Double file formats (in addition to the original U16).
-Reader is much cleaner. All byte type issues are abstracted out with Import/ExportQuantum pixels
-Reader handles all integer formats correctly, with accuracy limited by quantum depth.
What needs work:
-Reading floating point and double formats may need a little more attention.
-I would like to add IPL support to the test suite. (rwfile) Still trying to track down documentation/discourse posts on how to do this.

Coder has been benchmarked with 6.3.5 Q32.

ftp://depot.aecom.yu.edu/pub/b8GM3S012I6mGI6c059/ipl.c

Test images are at:
ftp://depot.aecom.yu.edu/pub/b8GM3S012I ... 9/8bmf.ipl
ftp://depot.aecom.yu.edu/pub/b8GM3S012I ... /16bmf.ipl
ftp://depot.aecom.yu.edu/pub/b8GM3S012I ... 9/8bsf.ipl
ftp://depot.aecom.yu.edu/pub/b8GM3S012I ... 916bsf/ipl

Best,

Sean

[magick]

We'll get your updated coder in ImageMagick 6.3.5-7 sometime tomorrow. Thanks.

[seanburke1979]

Excellent! Thanks.

Sean

[seanburke1979]

Looks like it didn't get into 6.3.5-7 or 8. Our public ftp server only holds accounts for one week (don't ask). Here is an updated link:
ftp://depot.aecom.yu.edu/pub/UNm0FMt06357hE9MO986/ipl.c

As always, drop me a line if there any questions or problems.

Sean

[magick]

We got your update IPL coder into ImageMagick 6.3.5-9 Beta a week or so ago. See ftp://ftp.imagemagick.org/pub/ImageMagick/beta .

[seanburke1979]

Ahh, ok. I wasn't aware that there was a separate beta release. Sorry for the confusion.

Sean

[anthony]

seanburke1979, is your coder now in place?
Does it save HDRI floating point values correctly.

I am wanting to add it to the list of image formats that are HDRI compatible in IM Examples section on HDRI
http://www.imagemagick.org/Usage/basics/#hdri

[seanburke1979]

Its funny that you bring that up - I was just working on a Fourier Transform routine for IM. HDRI would be a total necessity for any fourier work. Do you have an example of a floating point HDRI image? If so, you could just convert it from MIFF (for example) to IPL and then back again.
As an aside, I have tested both the float and double types, but only with my own images since I have had such a hard time finding stock HDRI images. A collection of benchmark images would help anyone who is writing a coder.

Best,

Sean

[magick]

Go to the OpenEXR site. They have tens of sample HDRI images.

[anthony]

Its funny that you bring that up - I was just working on a Fourier Transform routine for IM. HDRI would be a total necessity for any fourier work.

Oh that would be good for large scale convolutions, specifically finding sub-images in images as well as other types of image cleaning work. Look forward to it.

though I have seen examples of fourier transformed images using just the 0.0 to 1.0 range. I believe they make 'gray' the 'zero' position, or something like that.

Talk to fred Weinhaus. he would be happy to give you lots of advise on this type of thing. he's been most helpful in my Filter and Resampling work for distortions.

[seanburke1979]

Sounds good. How do I go about contacting him? I have some working code, but I'm not sure it would pass a security audit. Maybe he can give me some pointers.

Best,

Sean