顔面熱画像によるドライバーの一過性覚醒度の推定

概要

論文の詳細を見る
This study is aimed for the development of the quantitative measurement technology of the transient awakening degree fall based on a facial skin temperature change. Generally the characteristic change that accepted sleepiness levels such as eye blink rhythm or the movement of lips by an awakening degree fall appears in the face expression. On the other hand, the facial skin temperature changes by the action of the autonomous nervous system with the awakening degree fall significantly. In this study, we demanded the relevance that there was between this face skin temperature change and a face expression change. And we tried the construction of the model that estimated awakening degree based on a change pattern of the time and space of the facial skin temperature. We used a hierarchical model neural network for the modeling. We learned an explanation variable, face expression evaluation value as a purpose variable by the change pattern of the time and space of the facial skin thermal image. Concretely Speaking, we measured a heat picture, the number of cardiac beats, and facial expression evaluation. And we presumed the facial expression evaluation value from the facial skin temperature using the model. As a result of the experiment, the proposed model showed a possibility that it could presume in the stage of 1 to 3 of a facial expression evaluation value.