Browse Prior Art Database

A Method for Speech Recognizer Input to Computer Applications in a Windowing System

IP.com Disclosure Number: IPCOM000104079D
Original Publication Date: 1993-Mar-01
Included in the Prior Art Database: 2005-Mar-18
Document File: 2 page(s) / 77K

Publishing Venue

IBM

Related People

Andreshak, JC: AUTHOR [+4]

Abstract

Disclosed is a method for using speech recognition input for applications in a computer windowing system.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

A Method for Speech Recognizer Input to Computer Applications in a Windowing System

      Disclosed is a method for using speech recognition input for
applications in a computer windowing system.

      Automatic speech recognition (ASR) systems translate spoken
speech into text or commands.  This recognized text can be sent as
input into computer applications in place of keyboard or mouse input.
Two problems arise when an ASR system is used for application input
where multiple applications are running in a computer windowing
system such as X-Windows or OS/2 Presentation Manager*.

      One problem is that the ASR system must determine to which
application recognized text and commands should be sent as input.
Several different applications may be running concurrently in
different windows of the computer operating system, but a user
usually desires input to only one application at any given moment.
Several instances of the same application may also be running, with
the user interacting with only one instance at a time.

      Secondly, the ASR system must determine what text or command
input should be generated when a user speaks a word or phrase.  The
set of allowable user input is different for different computer
applications.  If an application's input is coming from an ASR
system, then text or commands generated by that system (the "active
vocabulary") must be restricted to that which is acceptable input to
the application.  However, if a user is running multiple applications
concurrently, then the ASR system must determine how to vary the
active vocabulary as different applications receive ASR input.

      This article introduces the concept of "speech focus", and
having an ASR text and command router which can control this focus.
In much the same way that keyboard input in computer windowing
systems has a concept of input focus, the speech focus determines
which application among those being run concurrently should receive
input from an ASR system.

      Just...