Supporting creative product/commercial design with computer-based image retrieval

  • Authors:
  • S. J. Westerman;S. Kaur

  • Affiliations:
  • University of Leeds, Leeds, UK;University of Leeds, Leeds, UK

  • Venue:
  • Proceedings of the 14th European conference on Cognitive ergonomics: invent! explore!
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Motivation -- The task of creative product/commercial design can be supported by the retrieval of inspirational images from computer databases. However, 'traditional' computer-based information retrieval algorithms, that emphasise convergent over divergent processes, may not be well suited to this task context. The reported empirical study examines related issues, including effects of users' search strategies. Research approach -- Participants (Engineering and Design students: n=12) were given a car design brief and asked to perform five searches for inspirational images. Participants rated the first 10 retrieved images for a variety of constructs including inspirational value and semantic distance to search terms. Findings/Design -- Inspirational images tended to be those that were semantically proximate to the search terms used to retrieve them. However, there were sufficient exceptions to suggest that semantic/conceptual diversity also plays an important role. Some key image attributes, relating to inspirational value, were identified and considered with regard to supporting effective task performance. Finally, specific search strategies were found to be associated with the retrieval relatively more inspirational images, although effect sizes were small. Take away message -- Creative design tasks require information systems support for both convergent and divergent processes. Improvements can be made in terms of the design of interface, retrieval algorithms, and training of users' search strategies.