Automatic recording of meal patterns using conductive chopsticks

Dietary education continues to attract attention. One target of such attention is the meals consumed alone by senior citizens whose eating habits are often unhealthy, inadequate, and irregular. To solve this problem, the life-log data of eating habits are effective. However, conventional sensing approaches with video cameras suffer from privacy concerns and low accuracy from poor lighting conditions. This paper proposes to solve this problem with a new method that records an ingestion log without burdening users. Our proposed method, which focuses on the closed circuit conductivity that is constructed from food, chopsticks, and a human body, determines the resistance value of the closed circuit and detects two actions: grasping food and carrying it to the mouth. We implemented our prototype chopsticks and experimentally verified our proposed methods. The accuracy of carrying food to the mouth exceeded 99% using F-measure. Our evaluation also clarified that our proposed method identifies participants who fail to join meal communities around a table.