Social Groups Detection in Crowd by Using Fuzzy Automatic Clustering with PSO

Detecting social groups is one of the most important and complex problems which has been concerned recently. This process and relation between members in groups are necessary for human-like robots in near future. Moving in the group means to be a subsystem in the group, in other words, a group containing two or more people can be considered in terms of same direction of movement and speed of the movement. All databases have some information about trajectories and labels of the members. The aim is to detect social groups containing two people or more or detecting individual motion of a person. For detecting social groups in the proposed method, fuzzy automatic clustering with PSO is used. At first, the locations of all people in frequent frames are detected and the average of locations is given to fuzzy automatic clustering with PSO. The proposed method in valid databases provides reliable results and is compared to state-of-the-art methods. A method providing better results compared to the proposed method needs all training data to be considered for training step but the proposed method is not required to be trained at all, and so this characteristic increases the ability of implementation for the robots.