CS313 PROJECT3
FACE-FINDER FOR EVENT PHOTOS

You are TO WORK INDIVIDUALLY.

   November 2004
 S  M Tu  W Th  F  S
14 15 16 17 18 19 20
         ^assigned
21 22 23 24 25 26 27
28 29 30  1  2  3  4
	  ^neural net training:  offline and instream
 5  6  7  8  9 10 11
	  ^intellectual history of AI
                ^due (i will start checking for code/blogs on this date!)


This project is a classic and is easy to describe.  All you need to do is
read in an image, e.g., http://wolf.cs.wustl.edu/~cs313/04-03/images, and
write it out so that ALL AND ONLY those pixels corresponding to a person's
skin have been rewritten as (255,0,0) pixels.  You just need to color all
of the skin red.  HA HA HA HA HA.  Good luck.

If you want an A, you need to put a box around ALL AND ONLY the faces
that are in the picture.

You may NOT use a skin-detector or face-detector that is already available
online.  In fact, you should write all of your vision code from scratch.
Don't even use an edge detector if you find one.  Write your own.
YOU MAY use any machine learning program you find online, if you choose to.

I would avoid duplicating the images directory if I didn't have to, in
order to save space.  You may want to back up your work, e.g., on k9, as
wolf has not convinced me of its stability yet.  It is however a fast
machine and you are welcome to crush it during this project.

I strongly recommend using C or C++ on this project if you use
the large versions of images.  On the other hand, I strongly
recommend you use the small versions of images.  You can even resample
to make them smaller.  If you do, then I recommend you use whatever
language you like the most.  In fact, you know what cause the segfault
error in class when we were trying to do an edge detector?  My failure
to distinguish char from unsigned char.  Use a scripting language on
a small image and be happy.

You might want to look at my unrawc.c, which you can use in this sequence:

	djpeg -pnm images/p0002512.jpg | unrawc | myfacedetector | cjpeg > myoutput.jpg

As usual, a good blog (in wolf/~cs313/04-03) will help you get an A.