Detection of Human Motion Gestures Using Machine Learning for Actual Emergency Situations
Ryohei Niiuchi, Hyunho Kang, Keiichi Iwamura · 2018
Wearable devices are widely used in daily lives. Hence, it is expected that wearable devices will be used in nursing care for elderly patients. However, a very rapid response is necessary under various emergency situations. It is necessary to contact a family guardian to quickly care for elderly family members. This study focuses on the detection of human motion gestures based on an acceleration sensor. Specifically, the study aim involves recognizing arbitrary hand gestures. Sufficient accuracy was obtained in the case of the Hiragana dataset. It is expected that the proposed approach offers a scalable technique for actual emergency situations.