Abstract:
This thesis proposes a method for controlling a computer via speech input that covers all key computer operations except for typing-based input with a small set of voice command. The method consists of multiple states, including: control method selection state, target selection state for changing the position of the cursor or selected item, and option state for applying additional operations to the target. The operations are clicking, double-clicking, right-clicking, scrolling, dragging, dropping and activating selected item. All are controlled by hum sounds with four different pitch types and a fricative sound. Hum sounds are dedicated to movement while the fricative sound is for stopping or activation. We have found that the proposed method showed better precision than Vocal Joystick while it was less tiring to use despite that longer intervals were usually needed to complete given tasks.