Files:

set2seed3.bst

BWRes3.rgb
ColRes3.rgb

results of running gp on the second(easier set)


input.file - evalution parameters

app.c - function set




Training set algorithm:


	There are 4 categories for points chosen:
	
	1.edge - point is in the center of the edge(all points around it are
	also edge points)
	
	2.border edge - pixel is an edge but some points around it are not
	edges
	
	3.not edge -  point doesnt have any adjacent edges to it
	
	4.close to edge -  point is not an edge but has edge point adjacent to
	it
	
	
User can specify number of each kind of points to be used(see input.file) also
training set can be renewed(new random points chosen) between generations(see
input.file).

Distribution used(appears in input.file)
	1. 700
	2. 700
	3. 800
	4. 700
	
Choice is done as follows:


	1.Table is created where each pixel is identified to belong to one of 4
	categories.
	
	
	2.on random basis points are chose and added into the table for each
	category separately
	
	3.if specified step 2 is done at regular intervals during the evalution
	
	



Fitness:
1-(correct_edge_identified/total_edges)*(correct_non_edge/number_nonedges)


Unused parameter: MaxPlane.rgb



Parameters:

1. GradientCrossed

2. MaxPosCrossed

3. MaxCrossed 	Red

4. 		Green

5. 		Blue

6. MinPosPlane

7. MinPlane	Red

8. 		Greeen

9. 		Blue