Olfoto: Designing a Smell-Based Interaction

Photograph of Stephen Brewster

Stephen Brewster

(University of Glasgow)

I will present a study into the use of smell for human-computer interaction, and in particular for searching digital photo collections. Many people now have large photo libraries on their computers and effective search tools are needed. Smell has a strong link to memory and emotion so may be a good way to cue recall when searching. Our study compared text and smell based tagging. For the first stage we generated a set of smell and tag names from user descriptions of photos, participants then used these to tag photos, returning two weeks later to answer questions on their photos. Results showed that participants could tag effectively with text labels, as this is a common and familiar task. Performance with smells was lower but participants performed significantly above chance, with some participants using smells well. This suggests that smell has potential. Results also showed that some smells were consistently identified and useful, but some were not and highlighted issues with smell delivery devices. I will also highlight some practical issues of using smell for interaction.



Bio

Stephen brewster is an EPSRC Advanced Research Fellow and leads the Multimodal Interaction Group at Glasgow, investigating how we can use the different senses to improve interaction with computers.


Tuesday 9th May 2006, 14:00
Robert Recorde Room
Department of Computer Science