<html><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Hi, <div> I just noticed that I should also have replied to this one.</div><div>Lassi's email summarizes the meat of what there was to say.</div><div><br></div><div>What the method does is basically to find the thresholds (by performing an exhaustive search on the image histogram) that minimize the between-class variance, which is (in latex notation - pdf render attached for convenience)</div><div><br></div><div>\sigma^2_b = \sum_{i=1}^{N_c}{\omega_i (\mu_i - \mu)^2</div><div><br></div><div>where \omega_i is the class proportion, \mu_i is the class mean intensity and \mu is the overall mean intensity.</div><div><br></div><div>I hope this helps, let me know if you need more details<br><div><br></div><div>Luca</div><div><br></div><div></div></div></body></html>