2010-2011-Voice Control Computing

Voice Control Computing Project

by Wytze Hazenberg

The aim of this project is to control the keyboard and mouse of a computer using only one's voice. The project was built in Java using the CMU Sphinx Toolkit (http://cmusphinx.sourceforge.net/). Sphinx matches verbal input to a pre-defined vocabulary.

The project's mouse interaction consists of an overlay containing nine blocks containing numbers (1 to 9). When a number is spoken, the overlay zooms in on that part of the overlay and the mouse moves to the specified block. Zooming can be repeated until the user calls one of the click commands.

A large vocabulary is important for the keyboard part. The spoken words are directly passed through to the active window. To gain more words one can expand the vocabulary using a recording functionality for filling the vocabulary.