A complex network-based approach to estimating the number of people in video surveillance

The estimation of number of people is an important research topic in video surveillance and processing applications. This paper presents a new approach to estimating the number of people in crowd scenes. A complex network-based algorithm is used to detect interest points and extract the global texture features in scenarios. We can obtain degree matrices and statistical measures which are highly correlated with the information of moving people. The indirect approach count moving people by establishing a mapping between the feature of moving interest points and the number of people of a crowd scene. Simulation results are also demonstrated to illustrate the effectiveness of our proposed method.