What is image segmentation?

Image segmentation is the partitioning of an image into distinct regions or categories that correspond to different objects or parts of objects. Each region contains pixels with similar attributes, and each pixel in an image is allocated to one of these categories. A good segmentation is typically one in which pixels in the same category have similar intensity values and form a connected region, whereas the neighboring pixels that are in different categories have dissimilar values. The goal of this is to simplify/change the representation of an image into something more meaningful and easier to analyze. 

If segmentation is done well, then all other stages in image analysis are made simpler. Hence, the quality and reliability of segmentation dictates whether an image analysis will be successful. But to partition an image into correct segments is often a very challenging problem.

Segmentation techniques can be either non-contextual (do not consider spatial relationships...