#VoHTK

VoHTK is a speech recognition web-application written in Go an powered by a speech recognition engine built with the HTK toolkit.

Check it hout here.

It uses the new HTML5 getUserMedia/Stream APIs for recording audio from a compatible browser.

The speech recognition engine is built using the HTK toolkit and based on Hidden Markov Models. It was originally trained using audio recordings from 630 speakers from the TIMIT speech corpus.

This project is based on my Master's Thesis work (Interactive Map with Voice Control for Mobile Devices) during my stay at the Technical University of Liberec in 2011.

###License MIT License