Segmentation

Image segmentation is the process of partitioning an image into non-overlapping regions of homogeneity, where homogeneity is defined in terms of an image feature such as intensity and certain textures. In the material image analysis, the segmentation task is a necessary pre-processing step to extract important information concerning material structures from images. Material researchers would typically like to first separate the foreground regions occupied by materials of interest from the remainder of an image and then conduct subsequent analyses on the segmented images, especially the material foreground. The subsequent analyses include, for instance, the morphology analysis for extracting the structures of the materials or the dispersion analysis for quantifying the spatial arrangement of certain material elements. In this chapter, we focus mainly on the major development in image segmentation for material images, as the subsequent image analyses after segmentation are to be discussed in the latter chapters. In the first section of this chapter, we describe the unique characteristics of material images and the challenges that they cause in image segmentation. In the second section, we present a sequence of analysis steps to address these challenges for material image segmentation. In the remaining sections, we describe different approaches that can be used to carry out the analysis steps in material image segmentation.