Stereo Vision, Grouping 3D points

hey guys, anyone know how to group 3D points to an object? 3D points was reconstructed from the disparity map built from two images from the same scene (stereo vision). Right now i just use the code from the demo given by typing stereo vision in matlab help browser. the one i'm asking is how to group this point, so that we can detect objects from images captured. For example if there is two objects in the scene (object I and object II), i want to group the points to 2 groups, one for object I and the other for object II. I had read some paper that can do this, but i am a total beginner in this matter, so anyone can give me some references or any help? I'll appreciate it very much. Thanks. *Sorry if my english not good.