No abstract

No abstract

Ensuring passengers' safety and comfort is a daily comfort for railway operators. In order to correctly size their transport plan, it is necessary to know in real time the number of users present in the trains. With the advent of artificial intelligence, several techniques now make it possible to detect and track passengers effectively. The combination of these 2 techniques can then make it possible to carry out a relevant count. In this article, we compare 2 counting methods based on the one hand on the detection of faces, on the other hand on the segmentation of shapes and their reidentification. We will show the promising results obtained, as well as the impact of physical phenomena such as occultation on the precision of the counting performed.