Feature Extraction & Image Processing, 2nd Edition

.

6

.

Flexible shape extraction

(snakes and other

techniques)

6.1 Overview

The previous chapter covered finding shapes by matching. This implies knowledge of a model

(mathematical or template) of the target shape (feature). The shape is fixed in that it is flexible

only in terms of the parameters that define the shape, or the parameters that define a template’s

appearance. Sometimes, however, it is not possible to model a shape with sufficient accuracy,

or to provide a template of the target as needed for the generalized Hough transform (GHT). It

might be that the exact shape is unknown or that the perturbation of that shape is impossible to

parameterize. In this case, we seek techniques that can evolve to the target solution, or adapt

their result to the data. This implies the use of flexible shape formulations. This chapter presents

four techniques that can be used to find flexible shapes in images. These are summarized in

Table 6.1 and can be distinguished by the matching functional used to indicate the extent of

Table 6.1 Overview of Chapter 6

Main topic Sub topics Main points

Deformable

templates

Template matching for deformable shapes.

Defining a way to analyse the best match.

Energy maximization, computational

considerations, optimization.

Active contours

and snakes

Finding shapes by evolving contours.

Discrete and continuous formulations.

Operational considerations and new active

contour approaches.

Energy minimization for curve

evolution. Greedy algorithm. Kass

snake. Parameterization; initialization

and performance. Gradient vector

field and level set approaches.

Shape

skeletonization

Notions of distance, skeletons and symmetry

and its measurement. Application of

symmetry detection by evidence gathering.

Performance factors.

Distance transform and shape

skeleton. Discrete symmetry operator.

Accumulating evidence of

symmetrical point arrangements.

Performance: speed and noise.

Active shape

models

Expressing shape variation by statistics.

Capturing shape variation within feature

extraction.

Active shape model. Active

appearance model. Principal

components analysis.

241