Quantification of perceptual design attributes using a crowd

DS 75-6: Proceedings of the 19th International Conference on Engineering Design (ICED13), Design for Harmonies, Vol.6: Design Information and Knowledge, Seoul, Korea, 19-22.08.2013

Year: 2013
Editor: Udo Lindemann, Srinivasan V, Yong Se Kim, Sang Won Lee, John Clarkson, Gaetano Cascini
Author: Ren, Yi; Burnap, Alex; Papalambros, Panos
Series: ICED
Institution: University of Michigan, United States of America
Page(s): 139-148
ISBN: 978-1-904670-49-0
ISSN: 2220-4334

Abstract

Crowdsourcing processes can be used for design concept creation and evaluation. They also provide opportunities to study and model quantitatively how humans deal with design problems. This paper explores the use of crowdsourcing to evaluate a perceptual design attribute and to create new design concepts using this attribute. As an example, we study how perceived automobile car safety can be modeled with respect to exterior car shape design using an efficient statistical learning algorithm. Experiments with subjects using Amazon's Mechanical Turk uncover several practical issues that must be addressed when applying machine learning methods to create safe-looking car designs using crowdsourced input.

Keywords: Preference learning, human-computer interaction, design crowdsourcing

Download