2017 IEEE/ACM International Conference on Connected Health: Applications, Systems and Engineering Technologies (CHASE) (2017)
Philadelphia, Pennsylvania, USA
July 17, 2017 to July 19, 2017
Human activity monitoring has become widely popular in recent years, and has been utilized in a vast number of fields and applications. Most of the activity recognition algorithms proposed have emphasized the use of inertial sensors in smartphone devices or other bodily-worn sensors. However, wearable inertial sensors are not interactive, and smartphones are not easily worn. Thus, with the advancement of smartwatches, unique opportunities exist to provide user interaction and highly accurate personalized activity recognition. Through the use of Active Learning, an interactive machine learning technique, specific behaviors can be learned by querying for unknown actions. This paper describes a smartwatch-based active learning method for activity recognition to identify 5 commonly performed daily activities. The results of this study revealed that this system can obtain a 93.3% accuracy across 12 participants. From our results, we demonstrate that an interactive learning approach using active learning in smartwatches has significant advantages over smartphones and other devices for activity recognition tasks.
Activity recognition, Monitoring, Uncertainty, Training, Acceleration, Learning systems
F. Shahmohammadi, A. Hosseini, C. E. King and M. Sarrafzadeh, "Smartwatch Based Activity Recognition Using Active Learning," 2017 IEEE/ACM 10th International Workshop on Cooperative and Human Aspects of Software Engineering (CHASE), Buenos Aires, Argentina, 2017, pp. 321-329.