Due to COVID-19, university lectures in Japan faced a change of face-toface lecture to a live online lecture. However, students at home need extra efforts to stay engaged to online lectures. This paper aims to increase engaged participants by changing the instructorߣs face shown over the lecture slide. First, students selected an in-favor and out-of-favor instructor photo which was both used to animate facial expressions in real-time using an open source deepfake software “Avatarify”. Students posted comments during the lecture and answered a simple quiz at the end. No significant difference was found for the quiz result, although the number of posting comments increased for the favorable photo. Therefore, we speculate that facial appearance influences the engagement of student participation in a live online lecture.