Digital Image Processing with C++

7.Segmentation

Segmentation consists in identifying and defining objects in the image: it is no longer a collection of pixels, but a set of disjointed regions forming a scene. Ideally, regions have a meaning (object, structure, signal, … or background) apart from their own characteristics (gray levels, shape).

Segmentation algorithms can be divided into two groups:

• region-based segmentation: pixels with common characteristics (called attributes or features), derived from the intensity of the pixels or from characteristics calculated on these intensities, are gathered together; the grouping criterion is evaluated in absolute (amplitudes) or in relative terms; decision rules (thresholds or metrics on the feature space) are then used to define ...