Authorities and security services have to deal with more and more data collected during events and on public places. Two reasons for that are the rising number of huge events, as well as the expanding coverage with CCTV cameras of areas within cities. Even the number of ground crew teams, that are equipped with mobile cameras, rises continuously. These examples show that modern surveillance and location monitoring systems come with need of suited assistance systems, which help the associated security workers to keep track of the situations. In this report, we present a first idea how such a system using modern machine learning algorithms could look like. Furthermore, a more detailed look on two state-of-the-art methods for human pose estimation is given. These algorithms are then investigated for their performance on the target domain of crowd surveillance scenarios using a small dataset called CrowdPose.
