Honours research

K.I.L.E.R

Retarded moron
Veteran
I'm getting myself into doing research, and I have been digging around for ideas in the field of computer graphics. So far I have found a few areas of research that I would like to do.
  1. Real time light scattering for translucent entities, such as smoke, haze, water, etc...
    I have read an IEEE paper that is dated Jan/Feb 2008 and it stated that real time scattering for translucent objects is still iffy (either too much memory used or too much computation).
  2. Ray tracing, this is an area of active development and Intel are pouring billions of dollars into it, I'm not exactly sure if I could make much of a contribution to this area given nearly all the best minds are on it already.
  3. Topology, rendering of geometry in the most efficient ways (i.e. Winged Edge data structure).
  4. Interpolation, I'm not even sure if any more work can be done in this area. I would like to be able to improve on computation costs (aka performance) in this area if I can.
I would like to be able get a few topics on hand and do some reading of my own before I go up to my lecturer so I can have an idea of what I want and we can have a discussion about it.
I'm still doing some reading from the IEEE but I'm having a lot of difficulty finding things.


Thanks.
 
The problem is that I want to do something that others will benefit from, even if it is just one person.
 
Topology and scene management would be my pick. So you could build virtual worlds with unlimited sizes, which you don't need to know all in advance.

Next would be a 3D internet, using that.
 
I've completely changed topics. lol
I'm doing "Time varying rates and immigration on population momentum". I'm going through a PhD's work and going to make improvements on it. :D
It's mathematics and numerical computing combined. At the end of this I hope to be more like Chalnoth in my understanding of math.

I'll be using Fortran for my work. :D
 
FORTRAN II, I presume? Because Fortran 2003 is just like most other modern languages.

I still think Perl is quite masochistic, or for something mind bending you might want to use Haskel or OcamL.

Then again, if you really want the best challenge, you should use basic LISP, without any of the massive default libraries. You can make them up if you go along, that's how they grew in the first place.

Consider: it's a completely syntax and value free language! The end result would be uniquely your personal language. And it's extremely suited for expert systems.
 
Why? Aren't all the numerical packages etc now available in, say, python?

True. The original program is written in Fortran, so I only have to make changes. The programming side of things wont be all that involved. I'm using Fortran 95.
 
Back
Top