Finally, i have finished the development of the basic recognicement system.
Now we have a powerfull recognicement HMM based system that makes no use of the floating point facilities.
That means that could be ported without problems to low power embedded systems.
You can download it from the "files" section
Today I released the DTW Module, now it has WIN32 support.
Take a look to the changelog to see the bugs corrected
I have started coding the first functions and structures related with the HMM Voice Recognition.
These are the centroid calculus.
You can see my evolution in the CVS:
This is the first release of a complete PC emulation of the future M68k speech recognicement system.
You can download it now!
The Hidden Markov Model voice recognition system to M68k embedded systems starts its developing.