The following demos of our experiments are available as Quicktime movies:

General MOT experiment (small)

General MOT experiment (large)

General MOT experiment (faster version)

Occluder demo MOT task with objects that disappear completely behind occluding bars without a decrease in tracking performance.
Virtual occluder demo Similar to previous demo, although occluding bars are not visible. This also has little effect on tracking performance.

Tracking and keeping track of items' identity. This demonstrates that it is very difficult to track and also to remember which item is which, as indicated by a number associated with each target

Implode/explode demo. This critical variant shows that an object must appear to pass behind an occluder for it to continue being tracked. When objects shrink (implode) and then emerge (explode) upon encountering virtual occluders, performance suffers. The only difference from the previous demo is that the object no longer behaves in a manner consistent with disappearance behind an occluding edge

The following demos show that not every well-defined set of features can be tracked. Simple squares are easily tracked. It gets harder when they are connected by a single line and much harder when they are connected by lines that make them into the outlines of square rods! Points are easy to track unless they happen to be the endpoints of a line.

More demos at Yale Vision Science: Attentional highbeams demo.