Social psychologists have been interested in identifying key factors that affect the efficiency of a group. In a group interaction, the leader is the person who has the authority, dominance and power to direct the people towards decision making. Therefore, the leader is the person who mainly affects the efficiency of the group. Nowadays, team leaders are being hired after being tested in multiple interviews each composed of different problem-solving tasks that allow observing the leadership skills of a person. Automatic detection of leaders using machine learning, computer vision and signal processing techniques is important as it is much more efficient and effective enough.
We are currently interested in automatically detecting emergent leaders (the leader who naturally arises from an interacting group rather than from a higher authority) from audio-visual recordings of group interactions. We investigate the efficiency of different machine learning methods to identify the emergent leaders in a meeting environment using nonverbal features. The nonverbal features are automatically extracted from a new dataset that is available for research purposes and can be downloaded from [here] .