Networked humanoid animation driven by human voice using extensible 3D (X3D), H-Anim and JAVA speech open standards

Abstract

Speech-recognition technology is beginning to be used in automobiles, telephones, personal digital assistants (PDAs), medical records, e-commerce, text dictation and editing. Speech recognition can also be integrated into Virtual Environments (VEs) to create responsive virtual entities. Like the mouse, keyboard, and the trackball, Speech-recognition technology can enhance the control of a computer and improve communication. Dramatically expanding interest in the Internet and VEs has been gated by limited interactivity with human-avatar models. As more users begin interacting with avatars in VEs, designers are prompted to create more realistic, humanlike avatars. This quest for realism needs to go beyond visual aspects to include speechrecognition technology, which can greatly augment the realism of these avatars. This thesis presents design and development of a Voice User Interface (VUI), which maps to a set of behavioral motions for humanoid avatars using Extensible 3D (X3D) graphics, the Virtual Reality Modeling Language (VRML), Humanoid Animation (H-Anim) Standard and Java Speech API. The VUI includes a suitable speech-recognition component for application-command vocabularies. This thesis also demonstrates interchangeability of both avatars and animation behaviors, and creates networked humanoid animation driven by a human voice.