Even four-month-old children identify objects due to their appearance and motion. They predict the re-appearance of objects after full occlusion and they can also distinguish between occlusion and disappearance for instance on the horizon. Motivated by these abilities, we built a framework which uses Gestalt principles to group lines to closed convex contours (closures) and track these closures. In subsequent frames new closures are found and interest points are initialised to track the closures. Consistent disappearance of interest points marks an occlusion event. The occlusion front defines the boundary of the occluding objects. Furthermore the depth relation between occluding and occluded objects is deduced from the occlusion event. We want to demonstrate this in a video sequence where a hand empties a shopping basket. Objects in a basket are grasped and placed on a table. The objects interact several times, get occluded and re-appear again. At the end of the experiment, each object should be known, i.e. surrounded by a closure.