Group-Leader Tracker
An AI resource to track "group-leader" social interactions within videos

This asset includes motion triggers and spatial interactions as an artificial cognitive process to be endowed in a robotic assistant or intelligent vision system. It can track and reason about social grouping with highly defined roles, in order to collaborate and support humans dealing with large groups of people.
Execution environment: windows
Version of the environment: 7
Installation: Run the .exe file in a terminal as:
group_leader_tracker.exe path_to_your_video_file.avi
and observe how the target group is categorized in the roles of leader (blue box), group members (green box) and non-group members (red box). Direction estimation of the leader of the group is also shown with a blue arrow and it is also sensitive to magnitude. Motion detection is indicated with a yellow box.
More information: This resource is the outcome of the research held in a doctoral stay at the National Institute of Informatics (NII) as an active member of the IDEAI-UPC research center. Our aim was to track and reason about social grouping with highly defined roles using motion triggers and spatial interactions as an artificial cognitive process to be endowed in a robotic assitant or intelligent vision system, in order to collaborate and support humans dealing with large groups of people. This action not only empowers the human expert to focus on their tasks, but also serves to avoid accidents and missing persons through smart surveillance. It represents one of a wide range of possible applications and future scenarios where group interactions are a key aspect in ambient intelligence and for robots to understand and effectively participate in social environments.
Reference to: https://github.com/karla3jo/Group-LeaderTracker
This code is open-source and free for anyone to experiment further with it :-) If you publish any work that uses this software or refers to it in any way, please just cite us:
Trejo, K., Angulo, C., Satoh, S. I., & Bono, M. (2018). Towards robots reasoning about group behavior of museum visitors: Leader detection and group tracking. Journal of Ambient Intelligence and Smart Environments, 10(1), 3-19.