A novel scheme for the detection of object boundaries is presented. Th
e technique is based on active contours evolving in time according to
intrinsic geometric measures of the image. The evolving contours natur
ally split and merge, allowing the simultaneous detection of several o
bjects and both interior and exterior boundaries. The proposed approac
h is based on the relation between active contours and the computation
of geodesics or minimal distance curves. The minimal distance curve l
ays in a Riemannian space whose metric is defined by the image content
. This geodesic approach for object segmentation allows to connect cla
ssical ''snakes'' based on energy minimization and geometric active co
ntours based on the theory of curve evolution. Previous models of geom
etric active contours are improved, allowing stable boundary detection
when their gradients suffer from large variations, including gaps. Fo
rmal results concerning existence, uniqueness, stability, and correctn
ess of the evolution are presented as well. The scheme was implemented
using an efficient algorithm for curve evolution. Experimental result
s of applying the scheme to real images including objects with holes a
nd medical data imagery demonstrate its power. The results may be exte
nded to 3D object segmentation as well.