顔動画像を用いた主観的「おいしさ」定量計測手法に関する検討

抄録

This paper presents a novel method to measure subjective palatability using facial expression via video analysis. We recorded facial expression using 30fps video camera to classify them to each subjective palatability using 512 facial landmarks. Self-Organizing Maps (SOM) is applied to them as classification algorithm. In this paper, we verify the usefulness of this method to use for subjective palatability sensing. This result has a tendency such as symmetrical facial expressions if a subject feels good taste, and unsymmetrical facial expression if a subject feels bad taste. This indicates the potential to measure subjective palatability using this facial expression pattern. From these results, our proposed method is useful for subjective palatability sensing.