My name is Pedro Montiel, I am from Mexico City.
Recently I have started a scholar project involving "spoken word recognition".
What I want to do is to program an application for "voice command recognition". Basically train the system to recognize the word I am saying.
I have took a look to your MARF framework and tested your "SpeakerIdentApp", and start to take it as a base for what i want to do. The thing is that I gave for training .wav files (recorded with my voice) for five different words, and then ask the program to recognize what I said in a new .wav file. But I am not getting the same great results you achieved in identifying a speaker.
I think the problem is that, if I understood well your "report.pdf", that all my sound files have being recorded by the same person.
Can you give me some advise to use your MARF framework to achieve my goal????
Thanks in advance.