Perceptual image retrieval using eye movements

Abstract

This paper explores the feasibility of using an eye tracker as an image retrieval interface. A database of 1000 Corel images is used in the study and results are analysed using ANOVA. Results from participants performing image search tasks show that eye tracking data can be used to reach target images in fewer steps than by random selection. The effects of the intrinsic difficulty of finding images and the time allowed for successive selections were also considered. The results indicated evidence of the use of pre-attentive vision during visual search.

Keywords:

• Computer interface
• Image retrieval
• Visual attention
• Eye movements
• Search performance

Acknowledgements

The authors acknowledge the support of BT Research and Venturing, SIRA and the Engineering and Physical Sciences Research Council for this work. The work was conducted within the framework of the EC-funded Network of Excellence (MUSCLE) 20.