Multi-object tracking is a challenging computer vision task that is a research hotspot in the literature. Although current one-stage methods can jointly optimize detection and appearance embedding ...

Multi-object tracking (MOT) in cluttered and dynamic environments remains challenging, especially for maneuvering objects. While trajectory-based random finite set (RFS) filters provide principled ...