Advances in Robust Multimodal Interface Design

  • Authors:
  • Sharon Oviatt

  • Affiliations:
  • -

  • Venue:
  • IEEE Computer Graphics and Applications
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

A well-designed multimodal interface that fuses two or more information sources can be an effective means of substantially reducing recognition uncertainty. Robustness advantages in multimodal systems have been demonstrated for different modality combinations, for varied tasks, and in different usage environments. Perhaps most importantly, the error suppression achievable with a multimodal system, compared with a unimodal one, can be in excess of 40 percent. In addition to improving overall recognition rates, a multimodal interface can perform in a more stable and effective manner across a variety of challenging user groups and real-world settings. This article reviews recent demonstrations of enhanced robustness for three types of multimodal interfaces, including ones that process speech and pen, speech and lip movement, and multibiometrics (physiological and behavioral) inputs. It concludes by discussing general design strategies for optimizing the robustness of future multimodal systems.