Interactive Dance Lessons through Human Body Pose Estimation and Skeletal Topographies Matching
6 Pages Posted: 28 Mar 2019
Date Written: 2018
During elementary period of learning, children are enthusiastic and spontaneous. This is very intuitive and natural human behaviour. But, the growth of affordable mobile devices have given us the possibility to experience diverse learning without physical movement. It is a considerable threat towards their physical and mental development. In this work, we tried to explore the possibilities of natural human-movement interaction with machine through ExerLearning experience, that is, gamified learning through exercise. We device a learning game for teaching simple dance moves to children which encourages them to perform a pose to match the given dance move. The findings of this work are very encouraging for stimulating students to perform exercise while learning, which effectively shortens the learning time with increased enthusiasm toward the content. Though ExerLearning approaches have been in practice with special devices, our stress on work was cost effective skeletal movement design, independent of body markers and special devices. We attempt to replace conventional skeleton tracking, provided by devices such as Microsoft Kinect, with a single camera, leveraging the power of Deep Learning to perform the same task. The anecdotal description and quantitative data reveals the effect of the enhanced learning outcome on the participants. The design recommendations base on the experimental data can pave a new paradigm of interactive and cost effective ExerLearning.
Suggested Citation: Suggested Citation