Split an audio file using voice recognition

Having an input audio file, representing an interview where 2-4 persons talk, I would like to identify automatically who talks at some point and split the file into small audio files stored in different folders, one for each person.

P1...5: Person 1...5
Original Audio File: [P1][P2][P1][P2][P3][P2][P1]
                     1   2   3   4   5   6   7

Output folders:
  P1
    1.mp3
    3.mp3
    7.mp3
  P2
    2.mp3
    4.mp3
    6.mp3
  P3
    5.mp3

What kind of programs would do such an identification and split the original file automatically?

It can be anything that works locally: for Windows, Mac or Linux.
Free and open-source software would be preffered.