Hello, I am an undergraduate student in final third year, interested
in working on this project for GSOC.
I have few good personal projects such as AI soccer simulation in C#
where two teams, five players each, play against each other. I have
programmed several 2D and 3D games in engines/frameworks such as XNA,
Unity etc. where I implemented finite state machine agents, goal-based
agents, fuzzy logic, steering behaviors, pathfinding, decision-making,
etc. I have also written from scratch my own C++ game engine for a 2D
game. Other projects include implementation of lots of computational
geometry, image processing and numerical linear algebra algorithms.
I was very excited when I read that Larry Gritz, the primary GSOC
mentor, had a leading role on projects such as RenderMan. I have used
it before in Maya. My MEL programming demo reel for Maya was featured
on cgtalk.com frontpage in 2007.
Here is what I implemented and tested these days for this project, 800
lines of code(not counting test files):
1. Histograms
- histograms for 8-bit and 16-bit grayscale images
- luminance histograms and per channel histograms for rgb images
- cumulative histograms
2. Point operations for 8-bit grayscale images
- contrast, brightness, invert, threshold
- auto-contrast and modified auto-contrast (better handles extreme values)
- histogram equalization
- two different methods for histogram specification(matching one
image to another)
- gamma correction
- alpha blending
3. Filters
- box filter
- general linear filter of arbitrary size
- min and max non-linear filters
- median filter
4. Edges
- sobel and prewitt edge detectors
- laplace sharpening
- unsharp mask
My code is in oiiotool.cpp from line 1497 to the end, you can download it here:
http://www.mediafire.com/?65z3noreznzjwie
Put the ip and images folders in "path", change "path" variable on
line 1497 in oiiotool.cpp.
The images folder contains test images used as input for the
algorithms. The ip folder contains output images from the algorithms
and test files for the above 4 topics. The tests are just code for
main in oiiotool.cpp. I commented the default code in main to test the
algorithms. I have not had the time yet to make it possible for users
to call them from command line.
I was not sure if I should submit my code to github, so I decided to
send it here and get some comments back. This is my first experience
on open-source project and I don't have much experience with github.
Finally, to summarize, my proposal for GSOC is lots of image
processing / computer vision algorithms and I would like to create a
Qt GUI for each. Let me know what you think.
Thank you, looking forward to hear from you guys.
Stefan
_______________________________________________
Oiio-dev mailing list
[email protected]
http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
