Facial expressions and gestures greatly affect the impression of speakers in interpersonal communication. In order to evaluate communication quantitatively, it is necessary to observe the behavior of scenes such as facial expression and face orientation in time series. However, the traditional assesment methods involve many subjective judgments. In this study, we propose a system to automatically analyze nonverbal communication using a vision camera. This system can absorb the difference in duration of time for the same movement (gesture). We will use the proposed system to improve interviewing and communication skills in the future.