hey all
i'm working on a project for continuous speech recognition.
i'm done with the speech processing - fft, mel spectral analysis and finally kmeans clustering.I'm using HMM's for the project and i'm using commands such as hmmtrain, hmmviterbi.oh and btw, i am not using any speech corpora if that makes sense.
my question is, given a model and a sequence, how can i find if there is a match when i input a new sequence? a formula or a module(an m file) would help.
Regards
Guru
i'm working on a project for continuous speech recognition.
i'm done with the speech processing - fft, mel spectral analysis and finally kmeans clustering.I'm using HMM's for the project and i'm using commands such as hmmtrain, hmmviterbi.oh and btw, i am not using any speech corpora if that makes sense.
my question is, given a model and a sequence, how can i find if there is a match when i input a new sequence? a formula or a module(an m file) would help.
Regards
Guru
