Robust Recognition of Reading Activity in Transit Using Wearable Electrooculography
In this work we analyse the eye movements of people in transit in an everyday environment using a wearable electrooculographic (EOG) system. We compare three approaches for continuous recognition of reading activities: a string matching algorithm which ex
- PDF / 2,301,699 Bytes
- 19 Pages / 430 x 660 pts Page_size
- 62 Downloads / 181 Views
2
ETH Zurich, Wearable Computing Laboratory [email protected] Lancaster University, Embedded Interactive Systems Group [email protected]
Abstract. In this work we analyse the eye movements of people in transit in an everyday environment using a wearable electrooculographic (EOG) system. We compare three approaches for continuous recognition of reading activities: a string matching algorithm which exploits typical characteristics of reading signals, such as saccades and fixations; and two variants of Hidden Markov Models (HMMs) - mixed Gaussian and discrete. The recognition algorithms are evaluated in an experiment performed with eight subjects reading freely chosen text without pictures while sitting at a desk, standing, walking indoors and outdoors, and riding a tram. A total dataset of roughly 6 hours was collected with reading activity accounting for about half of the time. We were able to detect reading activities over all subjects with a top recognition rate of 80.2% (71.0% recall, 11.6% false positives) using string matching. We show that EOG is a potentially robust technique for reading recognition across a number of typical daily situations.
1
Introduction
Activity recognition has recently emerged as a key area of research in building context-aware interfaces for mobile and pervasive computing. The problem of recognising physical activity in mobile situations, for example using body worn sensors, has been investigated by several researchers [17,21]. However, recognition of activities based on more subtle cues, such as user attention and intention - a far more difficult problem - remains relatively unexplored. A rich source of information on user activity is in the movement of the eyes. The paths that our eyes follow as we carry out specific activities also reveal much about the activities themselves. This is particularly true for activities with very specific eye movements, such as reading. Reading is a pervasive activity, e.g. on computer screens at work, advertisements and signs in public, and books read at home or while travelling. Thus information on a person’s reading activities can be a useful indicator of his daily situation as well as a gauge of task engagement and attention. Attentive user interfaces could comprise the current level of user interruptability or provide assistance to people with reading disabilities by automatically magnifying or explaining words or context in the text (for example see [15,10]). J. Indulska et al. (Eds.): Pervasive 2008, LNCS 5013, pp. 19–37, 2008. c Springer-Verlag Berlin Heidelberg 2008
20
A. Bulling et al.
We propose Electrooculography (EOG) as a novel measurement technique for recognition of user activity and attention in wearable settings. EOG, in contrast to well established vision-based eye tracking, is measured with body-worn sensors, and can be implemented as a wearable system. Although requiring facial skin contact, we believe EOG electrodes can be designed to be relatively unobtrusive, such as through integration into spectacles. A compact
Data Loading...
