Wearable Gaze Trackers: Mapping Visual Attention in 3D

Rasmus Jensen, Jonathan Dyssel Stets, Seidi Suurmets, Jesper Clement, Henrik Aanæs

Publikation: Bidrag til bog/antologi/rapportKonferencebidrag i proceedingsForskningpeer review

660 Downloads (Pure)

Abstract

The study of visual attention in humans relates to a wide range of areas such as: psychology, cognition, usability, and marketing. These studies have been limited to fixed setups with respondents sitting in front of a monitor mounted with a gaze tracking device. The introduction of wearable mobile gaze trackers allows respondents to move freely in any real world 3D environment, removing the previous restrictions.
In this paper we propose a novel approach for processing visual attention of respondents using mobile wearable gaze trackers in a 3D environment. The pipeline consists of 3 steps: modeling the 3D area-of-interest, positioning the gaze tracker in 3D space, and 3D mapping of visual attention.
The approach is general, but as a case study we created 3D heat maps of respondents visiting supermarket shelves as well as finding their in-store movement relative to these shelves. The method allows for analysis across multiple respondents and to distinguish between phases of in-store orientation (far away) and product recognition/selection (up close) based on distance to shelves.
OriginalsprogEngelsk
TitelScandinavian Conference on Image Analysis : 20th Scandinavian Conference, SCIA 2017 Tromsø, Norway, June 12–14, 2017 Proceedings, Part I
RedaktørerPuneet Sharma, Filippo Maria Bianchi
Antal sider11
UdgivelsesstedCham
ForlagSpringer
Publikationsdato2017
Sider66-76
ISBN (Trykt)9783319591254
ISBN (Elektronisk)9783319591261
DOI
StatusUdgivet - 2017
Begivenhed20th Scandinavian Conference on Image Analysis. SCIA 2017 - Tromsø, Norge
Varighed: 12 jun. 201714 jun. 2017
Konferencens nummer: 20
http://scia2017.org/

Konference

Konference20th Scandinavian Conference on Image Analysis. SCIA 2017
Nummer20
Land/OmrådeNorge
ByTromsø
Periode12/06/201714/06/2017
Internetadresse
NavnLecture Notes in Computer Science
Vol/bind10269
ISSN0302-9743

Citationsformater