How to code this in matlab

Given-Seed which is selected as initial seed is centre pixel of an RGB image
For grow formula we use intensity based similarity between seed and 8 neighbour pixels of the seed
The similarity index between two neighbourhood pixels of RGB intensities (x,y) and (x+i,y+j)is given by

Dist=sqrt(Dr+Dg+Db);
where
Dr=((f(x+i,y+j,1)-f(x,y,1))^2

Dg=((f(x+i,y+j,2)-f(x,y,2))^2

Db=((f(x+i,y+j,3)-f(x,y,3))^2

[Sean de Wolski]

What have you done so far?

[Poonam]

This is what i have done for HSV image
function d=dist(f,seed,cp8)
r=rgb2hsv(f);
[m,n]=size(r);
seed=[m/2,n/2];
Dh=(r(x+i,y+j,1)-r(x,y,1)).^2;
Ds=(r(x+i,y+j,2)-r(x,y,2)).^2;
Dv=(r(x+i,y+j,3)-r(x,y,3)).^2;
d=sqrt(Dh+Ds+Dv);

Answer 1

[Image Analyst]

Calculate delta E as in my demo http://www.mathworks.com/matlabcentral/fileexchange/31118-color-segmentation-by-delta-e-color-difference. Basically that's the color difference.

Does the value of the seed change from pixel to pixel (difficult), or is it always going to be the color that you selected as the first seed point (very easy)?

[Poonam]

seed point is going to be always the color which was selected as first seed point

[Image Analyst]

Then my delta E app is exactly what you want. Download it and run it. If you want, you can replace imfreehand() with ginput(1) so that you pick a single pixel rather than outlining a region.