We aim to understand image sequences the way humans do in order to detect critical situations before worse happens. So we improve 6D-Vision and downstream methods of image understanding continuously. Here we will present some relevant examples, in particular those which are