A new context: Screen to face distance

Most would agree: human vision is the most important of the five senses. Tragically many elderly people lose their vision due to incurable diseases which could have been avoided if diagnosed early enough. Fortunately some of these diseases can be diagnosed or at least have their symptoms detected with the use of simple tests. The use of smartphone or tablet applications have become common for these tests, so eye diseases can be detected early and even at home. However, none of the smartphone or tablet applications considers the screen to face distance of a person doing an eye test to be an important parameter for these sorts of tests. In this paper we present an algorithm to derive a new context: the smartphone user´s screen to face distance. Our algorithm utilizes the smartphone front camera and an eye detection algorithm. After initializing the algorithm with person specific values, the algorithm continuously measures the eye to eye distance to derive the user´s actual screen to face distance. We also present an investigation on the algorithm accuracy and speed, which shows: a smartphone based screen to face distance measurement is possible in the distance range from 19cm to 94cm with a maximum deviation of 2.1cm and at a rate of three distance measurements per second.