In this work we present a formal comparative over the computational
complexity of two edge detectors, in one hand the Canny Edges Detector
and on the other an edges detector based on Rough Sets Theory which
has been proved as very efficient from the results point of view, so
we are now interested in their performances, i.e., we have developed
an analysis from the computational point of view. To develop such
study we have used ROSA Analyser tool which generates the Labelled
Transition System, LTS, corresponding to a process specified in the
Markovian Process Algebra ROSA. ROSA Analyser takes as input the
syntactical representation of the process and once its syntactical
structure has been properly layered represented -internally- by the
tool, it applies the Operational Semantics of ROSA, so producing the
corresponding LTS, which shows all the possible behaviours of the
system we are interested in. A clear advantage, also from the
computational point of view, of the encoder using Rough Sets Theory
has been found.