transformed color space in matlab

The transformed color space (R’, G’, B’) is defined as follows:

R' = R-mean®/sd®,

G' = R-mean(G)/sd(G),

B' = R-mean(B)/sd(B), where mean and sd(standard deviation) of color channels C computed over whole image where C which belongs to (R,G,B).

I have to apply transformed color space to find the relationship b/w the pair of sampled points.

Suppose P(i)Cm are the values of color channel cm which belongs to (R’, G’, B’) of the sampled point cm and P(j)Cn are the values of color channel cn which belongs to (R’, G’, B’) of the sampled point cn where p(i)=(xi,yi) and p(j)=(xj,yj)

if P(i)Cm < P(j)Cn

then 1

 else 0 
where cm and cn can be (R',R')(R',G'),(R',B'),(G',G'),(G',B')(B',B')

[Image Analyst]

How are you specifying i and j? Are you using ginput()? Why do you want to compare the difference between one point in one color channel to a point at a different location in a different color channel?

[wa]

I have used following formula for finding out point p(x,y) :

for a = 1:8

theta = (2*pie/8)*a

radius = 4

location p (X) = cos (theta) radius + centreX

p (Y) = cos (theta) radius + centreY for each channel R G B

after this I have to compare same location value in each channel as described above in my question

[Image Analyst]

I'm sorry but I'm having trouble visualizing this. Please attach a diagram or picture.