I believe from common sense that video shape recognition problems (identifying shape of a moving object) is of natural interest in many real world situations. The process of identifying is understood as recognizing/finding one selected changing shape/signal among all the present shapes/signals. Ideally, it should be possible to track the shape.
I believe that this process (of recognising) might be computationally expensive (depends on the resolution). I would very much appreciate a survey level paper that summarises achievements so far.
It would be even better if the paper would point out to existing limits due to engineerical/hardware/technological restrictions. I am sorry that the question is a bit fuzzy.