Voice Features extraction from .wav file containing two speakers talking with each others

I have an idea and I am trying to apply it in a speaker recognition system project. how can I modify the project? and in which (.m) file should I change in order to apply my idea?

The idea is that, I want to extract features from .wav file containing two speakers talking with each others, but each speaker occupies a single sound channel, i.e., speaker A talks in left channel and speaker B,while responding to speaker A, talks in right channel. I want to find a way that extract the features for the two speakers in this .wav file.

Note: the speaker recognition system project is done by "raghu ram". It is available in File Exchange.

[Shadi Ayyad]

does any one have an idea about it?