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 ...

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 ...

Researchers have developed a new high-speed way to detect the location, size and category of multiple objects without acquiring images or requiring complex scene reconstruction. Because the new ...