ABSTRACT
This chapter focuses on image segmentation techniques, used to extract, or partition images into sets of objects of interest, so that an image is no longer a collection of pixels, but a set of disjointed regions forming a scene. The main goal is to have a new representation of an image into something that is more meaningful, so easier, to analyze. Two approaches are possible: either aggregate pixels inside a region which are similar in respect to their image features, or split pixels to find the segmented region boundaries because their image features are different.
We start by describing edge-based segmentation algorithms (active contours, geodesic level-sets), then we move on to region-based approaches and image partitioning. As in all chapters, clear and concise CImg codes are provided together with the corresponding illustrations.
