Image Segmentation : Pixel Classification Proccess

Hello, i am working in a project where i want to do an image segmantation. I am trying to separate the backround from the objects, so I made 2 pixel samples one for the backround and one for the foreground. I need help to manipulate these samples. The procces I have in mind is the following:

1. make a 3d histogram for each color model ( backround, foreground) and fit a gaussian distribution to each of them - any suggestions how to do it? ( functions etc...)

2. estimate the likelihood of each pixel for each distribution and decide in wich class the pixel belongs - also here any implemanting suggestions?

I am not really very familiar with these field, this is my first try. I am just trying to sort the info I found out there.

Any help will be very welcome...