Developing Face Emotion Tracker for Quantitative Evaluation of Care Effects

Abstract

In 2017, the Japanese government declared to start “scientific long-term care” as a national policy. In the practice of scientific long-term care, it is essential to assess the quality and effect of care services, since the caregivers must know if the care was effective for the target person. Currently, however, the effect of the long-term care has been evaluated by subjective observation and/or the questionnaire. Hence, it is difficult to justify the quality and effect as such the evidence encouraged in the scientific long-term care. To cope with the challenge, this paper proposes a novel system Face Emotion Tracker (FET) that evaluates the effect of care as a transition of emotions of a person under care. The proposed system can produce real-time data quantifying emotions of the target person under care, which is more objective and fine-grained clinical data compared to the conventional manual assessment sheets. We also implement a prototype and conduct experiments using the prototype.