Though the noise removal capability of multivariate median filters has been
carefully investigated, a comprehensive analysis of their complexity is st
ill missing. In this work, the complexity of the most commonly used multiva
riate median filters is thoroughly analyzed. For each filter theoretical re
sults are derived and validated against experimental data, proving that com
putational complexity depends mainly on the approach adopted to sort multiv
ariate samples. Algorithms based on marginal ordering are very fast, wherea
s the use of an ordering scheme based on the aggregate sum of distances lea
ds to very slow algorithms. An intermediate behavior is observed for filter
s relying on reduced ordering. A fast algorithm for the implementation of t
he vector median based on 1-norm is also described which significantly redu
ces the complexity of this filter. (C) 1998 Elsevier Science B.V. All right
s reserved.