Dissertation/Thesis Abstract

Harnessing the capacity of the human voice for fluidly controlling computer interfaces
by Harada, Susumu, Ph.D., University of Washington, 2010, 187; 3406108
Abstract (Summary)

The human voice still remains largely unexploited in the general computing domain as a primary or supplementary input modality despite significant advances in speech recognition technology over many years While voice input has a number of potential benefits, especially for people with motor impairments, one of the major limitations of current speech-based interaction methods is their inability to provide fluid and continuous input, akin to pointing devices such as the mouse. In this dissertation, I describe work that I have conducted as part of the Vocal Joystick project to harness the non-speech characteristics of human vocalization to enable such fluid hands-free control using the voice.

The Vocal Joystick engine upon which my work is based converts various non-speech vocal features such as volume, pitch, and vowel quality into continuous as well as discrete signals 100 times a second. I conducted a number of user studies to determine the performance characteristics of using such vocal input for various tasks including target selection, steering, and rapid discrete inputs. A key result shows that with only 10 hours of practice, users can approach the level of performance comparable to a manual joystick for pointing, and exceed the performance of an existing speech-based pointer control method.

I also present a number of concrete voice-driven applications that I built and evaluated to situate the study of non-speech vocal input in realistic contexts. VoiceDraw is a hands-free drawing program that enables the user to create free-hand-style drawings with dynamically-controlled stroke thickness and speed using only their voice. VoiceGame Controller rapidly transforms various vocal signals into keyboard and mouse signals making it possible to play conventional computer games hands-free. VoicePen augments digital stylus input with continuous voice input. Voice Controller seamlessly integrates voice-driven pointer control with conventional speech commands and dictation functionality, creating a powerful and practical hands-free input modality.

Indexing (document details)
Advisor: Landay, James A., Wobbrock, Jacob D.
Commitee:
School: University of Washington
School Location: United States -- Washington
Source: DAI-B 71/05, Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Computer Engineering
Keywords: Computer interfaces, Hands-free controls, Voice controls
Publication Number: 3406108
ISBN: 9781109727913
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest