K. Egiazarian et al., BOOLEAN DERIVATIVES, WEIGHTED CHOW PARAMETERS, AND SELECTION PROBABILITIES OF STACK FILTERS, IEEE transactions on signal processing, 44(7), 1996, pp. 1634-1641
In the present paper, the theory of Boolean derivatives, the activitie
s of the arguments of a BF, and Chow parameters are studied from the p
oint of view of their application in the statistical analysis of a cla
ss of nonlinear filters-stack filters, The connection between the part
ial derivatives of a positive BF (PBF) and the selection probabilities
of stack filters is established, The notions of the weighted activiti
es of the variables of the PBF and weighted Chow parameters are introd
uced for the analysis, the computation of the joint selection probabil
ity matrix, and the sample selection probability vector of a continuou
s stack filter, Spectral approaches to the selection probabilities of
stack filters are derived, In particular, spectral algorithms with com
putational complexity O(2(N)), where N is the number of input samples
within an input window, are given for the computation of sample select
ion probability vectors, The difference of the spectral algorithms pre
sented here from the nonspectral ones is that spectral algorithms are
universal, i.e., their complexities are independent of the PBF, which
is used as the base for stack filtering. They are also straightforward
to implement, and fast spectral transforms exist.