Eating progress estimation based on depth images for medication management support

Takuo Suzuki, Marina Nakagawa, Kunikazu Kobayashi

Abstract: The authors developed an image processing technique for estimating the progress of eating behavior of an elderly recipient. The image processing is for depth images and does not require any visual markers on plateware such as dishes. The authors' medication management support system can make a reminder at the correct time (e.g., after eating) depending on the estimated progress. In addition, a human support robot will be able to pass a dose of medicines to the recipient in advance by predicting the end of eating behavior.