The skeleton of a 2D shape is an important geometric structure in pattern analysis and computer vision. In this paper we study the skeleton of a 2D shape in a two-manifold
, based on a geodesic metric. We present a formal definition of the skeleton
for a shape
in
and show several properties that make
distinct from its Euclidean counterpart in
. We further prove that for a shape sequence
that converge to a shape
in
, the mapping
is lower semi-continuous. A direct application of this result is that we can use a set
of sample points to approximate the boundary of a 2D shape
in
, and the Voronoi diagram of
inside
gives a good approximation to the skeleton
. Examples of skeleton computation in topography and brain morphometry are illu