Sn. Kalitzin et al., A computational method for segmenting topological point-sets and application to image analysis, IEEE PATT A, 23(5), 2001, pp. 447-459
Citations number
18
Categorie Soggetti
AI Robotics and Automatic Control
Journal title
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
We propose a new computational method for segmenting topological subdimensi
onal point-sets in scalar images of arbitrary spatial dimensions. The techn
ique is based on calculating the homotopy class defined by the gradient vec
tor in a subdimensional neighborhood around every image point. This neighbo
rhood is defined as the linear envelope spawned over a given subdimensional
vector frame. In the simplest case where the rank of this frame is maximal
, we obtain a technique for localizing the critical points, i.e., extrema a
nd saddle points, We consider, in particular, the important case of frames
formed by an arbitrary number of the first largest by absolute value princi
pal directions of the Hessian. The method then segments positive and and ne
gative ridges as well as other types of critical surfaces of different dime
nsionalities. The signs of the eigenvalues associated to the principal dire
ctions provide a natural labeling of the critical subsets. The result, in g
eneral, is a constructive definition of a hierarchy of point-sets of differ
ent dimensionalities linked by inclusion relations. Because of its explicit
computational nature, the method gives a fast way to segment height ridges
or edges in different applications. The defined topological point-sets are
connected manifolds and, therefore, our method provides a tool for geometr
ical grouping using only local measurements. We have demonstrated the group
ing properties of our construction by presenting two different cases where
an extra image coordinate is introduced. In one of the examples, we conside
red the image analysis in the framework of the linear scale-space concept,
where the topological properties are gradually simplified through the scale
parameter. This scale parameter can be taken as an additional coordinate.
In the second example, a local orientation parameter was used for grouping
and segmenting elongated structures.