No abstract

No abstract

This paper describes a vision based pedestrian detection and tracking system which is able to count people in very crowded situations like escalator entrances in underground stations. The proposed system uses motion to compute regions of interest and prediction of movements, extracts shape information from the video frames to detect individuals, and applies texture features to recognize people. A search strategy creates trajectories and new pedestrian hypotheses and then filters and combines those into accurate counting events. We show that counting accuracies up to 98 % can be achieved.

In this article, we present a novel method for evaluating guidance systems using an immersive virtual environment in combination with a mobile eye tracking system. Accurate measurements of position, locomotion, viewing frustum, and gaze are captured in the virtual environment. They are applied to the projection of an attention map onto the virtual 3D environment for visualizing the fixation in the environment as well as the amount of time objects were fixated. To demonstrate the method's applicability, we conducted an experiment with 24 participants evaluating a guidance system of a large public infrastructure. The results show that our method allows for the creation of attention maps as well as for the identification of objects of interest based on eye tracking

Several models for simulation of pedestrian movement have been proposed in recent decades. These models are primarily used in the planning and evaluation of large pedestrian infrastructures, such as transportation hubs, with a focus to increase comfort and safety for pedestrians. Although the number of proposed simulation models is increasing at a fast pace, not much is known about the properties of calibration procedures or the transferability of the models estimated in one setting to other settings. This paper compares three calibration methods for a slightly adapted social force model. The main emphasis lies in the characteristics of the data-generation process and the information contained in the data sets. The sensitivity of the model parameters of the calibrated model were investigated, and the transferability of the model to different scenarios was tested. Results revealed that the quality of the data had a strong effect on the suitability of different calibration strategies and that the information content in the scene under investigation limited the transferability of the results to other scenarios. These results suggest that several data sets with different characteristics do not need to be included in the calibration process to achieve a model that performs well in a wider variety of settings.