how to get the probability value of a pixel during maximum likelihood classification in ERDAS imagine ?

[LAKSHMI]

I'm trying to do a fuzzy land cover classification using maximum likelihood classification. I need to get the probability of each pixel to fall in a particular class. Some equations have been formed based on gaussian distribution to calculate the fuzzy membership value corresponding to each pixel. To calculate this value each pixels are to be considered individually and substituted in the following equation: D = (X-M)^T * (cov)^-1 * (X-M) where X is a 4*1 matrix representing the DN for a particular pixel in 4 bands. I'm trying to do this using erdas imagine model maker, but failed to call each pixel separately. How can I do this ? Can I use Spatial modeler language to modify the model ? How can we initialize a pixel in "while loop" of spatial modeler language ?