Chewing and Swallowing Pattern Recognition Using Sound Information
Mai Takeuchi, Masafumi Nishida, Masafumi Nishimura · 2024
Adequate dietary and fluid intake are important for maintaining good health. For the elderly, inadequate fluid intake is especially dangerous for increasing dehydration risk. Therefore, a system that can automatically detect food and water intake is needed. Although various studies have been conducted on food behavior recognition, few have focused on swallowing food and water. In this study, we propose a method to detect chewing, food swallowing, and water swallowing by using meal sounds collected with a skin-contact microphone using deep neural network. Comparison experiments showed that combining multiple microphones improves event detection performance.