Personal Information:
Name: Mahmoud Zidan
E-mail address: mmmzeg@gmail.com
IRC: zidan @#YafaRay  

Idea to implement:

Synopsis and Goals:
Frequency analysis of the radiance field [1]enables new techniques and applications, including Monte Carlo ray tracing sampling. Currently no support for the sampling methods based on the frequency content in YafaRay. We will implement efficient sampling methods for YafaRay based on frequency analysis of the radiance function.

Detailed description:
We will use the photon mapping to perform frequency analysis for the radiance function in the 3D and FFTW [5] open source library to perform fast fourier transform. Our task will include developing the ground truth analysis for the frequency domain of the light transport. And we will also implement new image-space sampling algorithms as explained in [1] and [2] for both diffuse and glossy materials.

Workflow and Usability:
Our ground truth analysis for the light transport will help YafaRay developers in further research in this area. And the new image-space sampling algorithms will allow YafaRay users to use efficient and choose among new methods for sampling the radiance field.

Temptative schedule:
Present- April 20
    Understand the YafaRay engine
    Continue reading references [1][2][3][4]
April 21- May 20
    Perform the ground truth experimental for the frequency analysis of the radiance function using photon mapping and FFTW     
    Analyze the frequency content of the light transport
    Analyze the frequency content of the visibly function
    Analyze the frequency content of the cosine term
    Analyze the frequency content of the diffuse BRDF
May 21 - July 04:
    Coding image-space sampling algorithm for diffuse surfaces based on the frequency content as described in [1] [2].
July 05 - July 12:
    Mid-term evaluation.
July 13 - July 31:
    Coding image-space sampling algorithm for glossy materials using the frequency information in [1] [2].
August 01 - 15:
    Try to add other feature as depth of field [3] or motion blur[4].
August 16:
    Final evaluation.

Bio:
I’m Mahmoud Zidan, an M.Sc. student in computer science department, Ain-Shams university, Cairo, Egypt. My major is Computer Graphics and very recently I have implemented various spatial partioning data structures and Monte Carlo ray tracer on GPU [6][7][8], currently I’m implementing the global photon mapping algorithm on GPU [9].

I’ve a strong background on ray tracing and global illumination algorithms, previously I’ve implemented the a 3D modeling tool, ray tracer and photon mapping  algorithm as my senior thesis[13].

Programming skill:
I’ve strong experience in C/C++  (9 years), OpenGL, GLSL, CUDA , Matlab. My computer graphics background includes an undergraduate course in CG using OpenGL, and a postgraduate course in CG. But for keeping updated to CG research, I’ve studied online courses in computer graphics by Ravi R. & O’Brien (2009) [10], Pat Hanrahan (2008) [11], and Parallel GPU computing course in [12]. I’m also keeping updated with recent CG publications through Ke-sen Huang page. As for the CG books my list includes but not limited to Jensen’s book, physically based rendering book, Cg book. My experience in the frequency domain includes a study and implementation of basic frequency filters and bilateral filters in image processing course [13].

References:
[1] http://people.csail.mit.edu/fredo/PUBLI/Fourier/ 
[2] http://www.eecs.berkeley.edu/~ravir/papers/firstorder/index.html 
[3] http://artis.imag.fr/Publications/2009/SSDHS09/ 
[4] http://www.cs.columbia.edu/cg/mb/ 
[5] http://www.fftw.org/ 
[6] http://graphics.cs.uni-sb.de/fileadmin/cguds/papers/2009/kalojanov_hpg20...
[7] http://www.cs.unc.edu/~lauterb/GPUBVH/paper.pdf 
[8] http://www.kunzhou.net/2008/kdtree.pdf 
[9] http://www.kunzhou.net/2009/GPUGI.pdf 
[10] http://inst.eecs.berkeley.edu/~cs294-13/fa09 
[11] https://graphics.stanford.edu/wikis/cs348b-08 
[12] http://www.seas.upenn.edu/~cis665/index-2008.htm 
[13] http://www.mahmoudzidan.com