<< Chapter < Page Chapter >> Page >

Conclusions and Future Improvements:

The project yielded reasonably accurate results using our independent algorithm. Unfortunately, it is a severely limited method of speech recognition. Feature-matching might be improved with an optimal determination of the number of Gaussians for the GMM of each sample, and recognition of female voices might be resolved easily enough with an appropriate frequency normalization filter, but the other aforementioned issues present much more serious problems, due to the sheer number of possible variations.

A more robust speech recognition system would certainly include a larger sample size, as well as recognize based upon phonemes instead of words, as many of the modern HMM-based systems do.

Given some more time, we would have tried to salvage our attempt at the LPC as well, this time using a support vector machine for improved feature-matching.

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Elec 301 project: voice recognition. OpenStax CNX. Dec 19, 2011 Download for free at http://cnx.org/content/col11396/1.3
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Elec 301 project: voice recognition' conversation and receive update notifications?

Ask