Hit-or-miss and Skeletons

Abstract

In contrast to morphological transformations described so far, the hit-or-miss transformation involves SEs composed of two sets. The first has to fit the object under study while the second has to miss it. Hence, the name fit-and-miss would have been more appropriate. Hit-or-miss transformations extract all image pixels satisfying a given neighbourhood configuration such as that corresponding to an isolated background or foreground pixel. Adding to an image all pixels having a given configuration leads to the thickening operator while subtracting them from the image defines the thinning operator.