Visual recognition of feeding robot

In recent years, due to the progress of aging progress, the number of Certified Nursing Care persons is on an increasing trend, and it is conceivable that the lack of human resources of carers will become more severe. As for meal assistance, which is one of nursing care services, it is thought that due to the shortage of personnel in the future, the assistance becomes sloppy and the care recipient becomes hard to enjoy the meal.
In this research, we aim to develop a robot that estimates comfortable pace distribution by using image processing and provides meals assistance. Specifically, we first use a camera that can measure the depth, and we will detect the chewing and swallowing of the subject when we are having a meal. We estimate eating pace by machine learning from these information. We control the robot to assist meal with estimated pace distribution.

Takanishi Lab, Waseda University
Specified nonprofit corporation wearable environmental information net promotion organization

◆Published literature
〇Domestic Conference
[1] Shyodai Yano,Atsuto Fujimoto,Takaaki Ohkawauchi and Jun Ohya,Study of Recognizing Eating Situations of Care Recipients by Analyzing RGB-D Sequences,Visual/Media Computing Conference 2017