Image Coding
Consider the occurrence of M gray levels in an image, with the
probability of occurrence of the ith gray level being
Prob (mi) =pi, i = 1, 2, . . . ,M.
A measure information h(p) should be a function of pi,
satisfying the following criteria:
h(p) should be continuous for 0 < p ə,
h(p) = 0 for p = 1 (event allways occurs -> no uncertainty),
h(p2) > h(p1) if p2 < p1,
if two independent image processes (or pixel) pi and qi are considered, h(pi, qi) = h(pi) + h(qi).
These requirements are met by h(p) = -logbp = logb1/p