Tampere University of Technology

TUTCRIS Research Portal

Enhanced gaze interaction using simple head gestures

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Details

Original languageEnglish
Title of host publicationUbiComp'12 - Proceedings of the 2012 ACM Conference on Ubiquitous Computing
Pages705-710
Number of pages6
Publication statusPublished - 2012
Publication typeA4 Article in a conference publication
Event14th International Conference on Ubiquitous Computing, UbiComp 2012 - Pittsburgh, PA, United States
Duration: 5 Sep 20128 Sep 2012

Conference

Conference14th International Conference on Ubiquitous Computing, UbiComp 2012
CountryUnited States
CityPittsburgh, PA
Period5/09/128/09/12

Abstract

We propose a combination of gaze pointing and head gestures for enhanced hands-free interaction. Instead of the traditional dwell-time selection method, we experimented with five simple head gestures: nodding, turning left/right, and tilting left/right. The gestures were detected from the eye-tracking data by a range-based algorithm, which was found accurate enough in recognizing nodding and leftdirected gestures. The gaze estimation accuracy did not noticeably suffer from the quick head motions. Participants pointed to nodding as the best gesture for occasional selections tasks and rated the other gestures as promising methods for navigation (turning) and functional mode switching (tilting). In general, dwell time works well for repeated tasks such as eye typing. However, considering multimodal games or transient interactions in pervasive and mobile environments, we believe a combination of gaze and head interaction could potentially provide a natural and more accurate interaction method.

ASJC Scopus subject areas

Keywords

  • Dwell time, Eye tracking, Head gestures, Selection