Speech-based navigation and error correction: a comprehensive comparison of two solutions

  • Authors:

  • Jinjuan Feng;Shaojian Zhu;Ruimin Hu;Andrew Sears

  • Affiliations:

  • Towson University, Computer and Information Sciences Department, 21252, Towson, MD, USA;UMBC, Interactive Systems Research Center, Information Systems Department, 1000 Hilltop Circle, 21250, Baltimore, MD, USA;Towson University, Computer and Information Sciences Department, 21252, Towson, MD, USA;UMBC, Interactive Systems Research Center, Information Systems Department, 1000 Hilltop Circle, 21250, Baltimore, MD, USA

  • Venue:
  • Universal Access in the Information Society
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Speech-based navigation and error correction can serve as a useful alternative for individuals with disabilities that hinder the use of a keyboard and mouse, but existing solutions available in commercial software are still error-prone and time-consuming. This paper discusses two studies conducted with the goal of improving speech-based navigation and error correction techniques. The first study was designed to improve understanding of an innovative speech-based navigation technique: anchor-based navigation. The second study was longitudinal, spanning seven trials, and was intended to provide insights regarding the efficacy of both traditional target/direction-based navigation and anchor-based navigation. Building on earlier studies that employed similar methodologies and interaction solutions, this paper also provides an informal evaluation of a new correction dialogue. Although the two solutions resulted in the same level of efficiency, the underlying strategies adopted were different, and the anchor-based solution allowed participants to generate better quality text and was perceived to be easier to use. These results suggest that the anchor-based solution could be a promising alternative, especially for novice users as they learn how to use speech-based dictation solutions. The findings of these studies need to be further validated with the involvement of users with disabilities.