This paper proposes a novel set of 16 features based on the statistics
of geometrical attributes of connected regions in a sequence of binar
y images obtained from a texture image. Systematic comparison using al
l the Brodatz textures shows that the new set achieves a higher correc
t classification rate than the well-known Statistical Gray Level Depen
dence Matrix method, the recently proposed Statistical Feature Matrix,
and Liu's features. The deterioration in performance with the increas
e in the number of textures in the set is less with the new SGF featur
es than with the other methods, indicating that SGF is capable of hand
ling a larger texture population, The new method's performance under a
dditive noise is also shown to be the best of the four.