Multidimensional Discrete Signals Description Using Rotation and Scale Invariant Pattern Spectrum

Abstract

Mathematical morphology was developed in the mid 1960’s by G.Matheron and J.Serra as a methodology for continuous and discrete multidimensional signal analysis. The basic idea underlaying this methodology is to trasform the original signal into a simpler and more expressive one, by interacting with a structuring element, called kernel, strategically chosen by the observer. A morphological operation is then constitucd by a transformation followed by some measurement on the transformed signal. The measurement on the transformed signal can be its lenght, area, volume, etc., depending on the dimension of the signal.