How to implement this algorithm ?

Question

Hello all,
can u help me in implementing the following algorithm in java?
1.suppose we start with a single pixel p and wish to expand from that seed pixel to fill a coherent region. lets define a similarity measure s(i,j) such that it produces a high result if pixels i and j are similar and a low one otherwise.
first, consider a pixel q adjacent to pixel p. we can add pixel q to pixel p's region if S(p,q)>T for some threshold T. we can the proceed to the othr neighbors of p and do likewise.The process will continue till all the pixels in the image are labelled or belong to one region or the other.

Answer 1

You may want to look at the various Connected components algorithms out
there. These (tend to) recursively search the neighbours of the seed
pixel and check for similarity. If the neighbours are similar to the
seed they are marked as being in the same region. These algorithms can in general be made to run using a stack and while loop rather then using recursion (which on large images can get you into memory trouble very rapidly).
A basic algorithm does the following
create a second image containing ints. This image is the output from the algorithm. different regions have different integer numbers assigned to them. Ensure all pixels have -1 as their value.
from your seed pixel check each of its neighbours in turn
// the output image of integers
private Image output;
// the input image to process
private Image input;
// origin of the image is in the top left
// assuming input image is greyscale and getPixel() / setPixel()
// return Pixel objects. regionvalue should be greater then 0
public void processPixel(int x, int y, Pixel compareTo, int regionValue)
   // am I similar to the pixel passed in
   // if so set my corresponding pixel in the output image
   if(isSimilar(input.getPixel(x,y),compareTo))
      output.getPixel(x,y).setGreyValue(regionValue);
   } else
      // i am not similar so bail out at let the caller check its
      // other neighbours
      return;
   //check north
   if( output.getPixel(x,y-1).greyValue() == -1)  // -1 means that its not been processed
      // this compares new pixel to this pixel
      processPixel(x,y-1,input.getPixel(x,y), regionValue);
   //check north east
  if( output.getPixel(x+1,y-1).greyValue() == -1)  // -1 means that its not been processed
      // this compares new pixel to the seed pixel
      // achives a different goal then the search for north given above
      // choose which goal suits you best
      processPixel(x+!,y-1,compareTo, regionValue);
   // check east
   //check south east
    //check north west
// nope you are not mistaken all 8 neighbours may need to be checked if
//  processing a greyscale image.
    return;
}matfud