How much do you read?: counting the number of words a user reads using electrooculography

We read to acquire knowledge. Reading is a common activity performed in transit and while sitting, for example during commuting to work or at home on the couch. Although reading is associated with high vocabulary skills and even with increased critical thinking, we still know very little about effective reading habits. In this paper, we argue that the first step to understanding reading habits in real life we need to quantify them with affordable and unobtrusive technology. Towards this goal, we present a system to track how many words a user reads using electrooculography sensors. Compared to previous work, we use active electrodes with a novel on-body placement optimized for both integration into glasses (or head-worn eyewear etc) and for reading detection. Using this system, we present an algorithm capable of estimating the words read by a user, evaluate it in an user independent approach over experiments with 6 users over 4 different devices (8" and 9" tablet, paper, laptop screen). We achieve an error rate as low as 7% (based on eye motions alone) for the word count estimation (std = 0.5%).