This paper deals with a method for the classification of
objects described by planar curves in an image.
It is suggested that an important aspect of shape classification methods
is the nature and degree of the invariance to deformation of the shape.
One extreme of this spectrum is the statistical description of shape.
Such a method is presented here.
Using scale space to describe shape statistically
allows for an unusual form of object classification.
The scale space used is one based on curvature measurements
extracted by a modified ``regularization'' method.
It allows a curve to be represented as a set of descriptors at various scales.
The spatial correlation of these
descriptors is used to produce a statistical description of a contour which
has similarities to a large-scale texture measure. The ``texture''
being measured is, in fact, the combination of substructures
that define the object's shape.