What's an appropriate rule of thumb for dealing with "magic numbers" (empirical parameters) in the development of an image analysis algorithm?
(A) Magic numbers are simply not allowed.
(B) More magic numbers make the algorithm more reliable.
(C) The use of magic numbers should be limited and their sensitivity be tested.
(D) It is not necessary to make magic numbers adaptive to image resolution because they are supposed to be set arbitrarily.
(E) none of the above