Susumu Harada is a Ph.D grad in the Computer Science and Engineering department at University of Washington. He invited me over to see a system he built, which allows someone to paint pictures using a voice-controlled mouse. Though the practical applications of this include other uses, drawing is the most intricate application because it involves lots of curves and mouse movements, which don’t lend themselves to other methods of accessibility.
The system uses a method of vowel sounds to move the mouse, and discrete constant sounds to click the mouse buttons. You can see more about the technology behind this at the Vocal Joystick Project.