We propose a pedestrian-pedestrian and pedestrian-vehicle interaction motion model for predicting the motion of the pedestrians and tracking pedestrians in traffic scene mixed with cars. Pedestrians’ motion in such scenes are often complex and highly correlated with both pedestrians and vehicles nearby. What’s more, vehicles and pedestrians running together will bring in long occlusions, what will make great challenge for tracking objects in long period. An interaction motion model combined with a path model is proposed to handle the sudden motion change and long occlusions. Experiments on real traffic scene sequences show the effectiveness of our method.
CITATION STYLE
Sheng, H., Liu, S., Ji, H., Chen, J., & Xiong, Z. (2014). A Pedestrian-pedestrian and pedestrian-vehicle interaction motion model for pedestrians tracking. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8887, pp. 270–280). Springer Verlag. https://doi.org/10.1007/978-3-319-14249-4_26
Mendeley helps you to discover research relevant for your work.