Multi-object tracking (MOT) lies at the heart of dynamic scene understanding in computer vision, with applications spanning autonomous vehicles, video surveillance, sports analytics and robotics. The ...

naturally. For example, a random object being considered may be a random connected graph, a random bounded-degree graph, or a random error-correcting code with good distance. A pseudo-random ...

Few-shot object detection addresses the challenge of recognising and localising novel object categories from only a handful of annotated examples. Traditional deep-learning detectors demand extensive ...