USAGE:


### Using the existing training set ###
1.Navigate to this directory in Matlab.
2.Run init, which will load in all images (s1-s20), masks (bw1-bw20), masked images(masked1-masked20),
	images minus masked portions (notmasked1-notmasked20), and compute the 3D histograms for target
	pixels (hist1) and non-target pixels(hist2)
3. Call comphist(<IMNAME>, <X>, <Y>, <S>, hist1, hist2), where <IMNAME> is any matlab image (e.g. one of s1-s20),
	<X> is the desired sliding window width, <Y> is the desired sliding window height, and S is the resolution
	in the window sliding amount in pixels. Assign the result to a variable.
	e.g. "res = comphist(s1, 4, 4, 2, hist1, hist2)"
4. Display the result with imshow.


### Make your own training sets ###
1. Do one of the following:
	1a. Create a new folder with a subfolder called "train", and copy all four .m files from this project into the former.
	1b. Delete or move all images in the "train" folder so that the train folder is empty
2. Collect a set of training images containing your objects of interest. Place them in the training folder and name
	them, in any order, "s1.jpg", "s2.jpg", "s3.jpg", etc. The numbering must start at 1 and must be continuous
	(e.g. you can't skip s2.jpg and just have s1.jpg and s3.jpg)
3. For every image "s<i>.jpg", where i is the index, run the following line:
	imwrite(roipoly(imread('train/s<i>.jpg')), 'train/s<i>mask.png')

	Every time you run this, the image will be loaded and you will be asked to draw a polygon around the object
	of interest in your photo. Select the points of the polygon by clicking, and close the polygon by double clicking
	anywhere or clicking on the first point. Double click the finished polygon to confirm. This creates a mask
	for this image with name "s<i>mask.png" in the training folder.
4. Go to step 2 in the existing training set section above.