Browse Prior Art Database

Interactive speech enrollment wizard accessible to screen readers

IP.com Disclosure Number: IPCOM000014947D
Original Publication Date: 2001-Oct-01
Included in the Prior Art Database: 2003-Jun-20
Document File: 2 page(s) / 99K

Publishing Venue

IBM

Abstract

Disclosed are a number of programmable methods for making the misrecognition point in a graphical user interface s interactive enrollment procedure available to a blind user. The Problem Many current speech dictation applications provide an interactive enrollment wizard. These interactive enrollment wizards present a panel of text to users for them to read for the purpose of collecting acoustic data with which to customize the recognition model. When a recognition problem occurs (either due to a system failure to recognize the user's speech or a user disfluency), these wizards leave all the text in the panel and indicate the resumption point with a visual indicator. The problem is that these visual indicators are not accessible to screen readers, which means that a blind user taking advantage of assistive technologies such as a screen reader cannot determine the correct point at which to resume reading. Solutions

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 77% of the total text.

Page 1 of 2

Interactive speech enrollment wizard accessible to screen readers

  Disclosed are a number of programmable methods for making the misrecognition point in a graphical user interface = s interactive enrollment procedure available to a blind user.

The Problem

Many current speech dictation applications provide an interactive enrollment wizard. These interactive enrollment wizards present a panel of text to users for them to read for the purpose of collecting acoustic data with which to customize the recognition model. When a recognition problem occurs (either due to a system failure to recognize the user's speech or a user disfluency), these wizards leave all the text in the panel and indicate the resumption point with a visual indicator. The problem is that these visual indicators are not accessible to screen readers, which means that a blind user taking advantage of assistive technologies such as a screen reader cannot determine the correct point at which to resume reading.

Solutions

There are several ways to address this problem. When the recognition problem occurs, the system could sound an alerting tone, followed optionally by a TTS or recorded voice presenting the remaining text to read on the panel. Another approach would be to delete the text that the user has successfully read, moving the problem text to the upper left corner of the panel, making it accessible through normal means to a screen reader. A third approach would be to hide the text that the user has...