A framework for object segmentation in vector-valued images is present
ed in this paper. The first scheme proposed is based on geometric acti
ve contours moving toward the objects to be detected in the vector-val
ued image. Object boundaries are obtained as geodesics or minimal weig
hted-distance curves, where the metric is given by a definition of edg
es in vector-valued data. The curve flow corresponding to the proposed
active contours holds formal existence, uniqueness, stability, and co
rrectness results. The scheme automatically handles changes in the def
orming curve topology. The technique is applicable, for example, to co
lor and texture images as well as multiscale representations. We then
present an extension of these vector active contours, proposing a poss
ible image flow for vector-valued image segmentation. The algorithm is
based on moving each one of the image level sets according to the pro
posed vector active contours. This extension also shows the relation b
etween active contours and a number of partial-differential-equations-
based image processing algorithms as anisotropic diffusion and shock f
ilters. (C) 1997 Academic Press.