We propose subtle foot-based gestures named foot plantar- based (FPB) gestures that are used with sock-placed pressure sensors. In this system, the user can control a computing device by changing his or her foot plantar distributions, e.g., pressing the floor with his/her toe. Because such foot movement is subtle, it is suitable for use especially in a public space such as a crowded train. In this study, we first conduct a guessability study to design a user- defined gesture set for interaction with a computing device. Then, we implement a gesture recognizer with a machine learning technique. To avoid unexpected gesture activations, we also collect foot plantar pressure patterns made during daily activities such as walking, as negative training data. Additionally, we evaluate the unobservability of FPB gestures by using crowdsourcing. Finally, we conclude with several applications to further illustrate the utility of FPB gestures.
Copyright (c) 2013 Koumei Fukahori