In this paper we address the problem of multi-object tracking in video sequences, with application to pedestrian tracking in a crowd. In this con- text, particle ¯lters provide a robust tracking framework under ambiguity conditions. The particle ¯lter technique is used in this work, but in order to reduce its computational complexity and increase its robustness, we propose to track the moving objects by generating hypotheses not in the image plan but on the top-view reconstruction of the scene. Compara- tive results on real video sequences show the advantage of our method for multi-object tracking.