Vassili A Kovalev 
, Maria Petrou 
and Yaroslav S Bondar
(1) Institute of Mathematics, Belarus Academy
of Sciences,
Kirova St., 32-A, Gomel 246652, Belarus
(2) School of Electronic Engineering,
Information Technology and Mathematics,
University of Surrey, Guildford GU2 5XH, UK
m.petrou@ee.surrey.ac.uk
A method is proposed for the characterisation of surface roughness of 3D surfaces. The method is based on the construction of a 2D co-occurrence matrix that expresses the frequency of occurrence of pairs of quantised values measuring chord length and relative orientation between the surface normals at the ends of the chord. The normalised matrices for various surfaces are compared either with the help of standard methods for comparing histograms, or with the help of features extracted from them. The approach is demonstrated with the help of an inventory of 3D medical images.