Medical image segmentation using openings and closings with reconstruction criteria

In this paper, we use a type of morphological filters, the openings and closings with reconstruction criteria, for image segmentation purposes. Those filters are a class of extended levelings, and they present interesting features for image processing and analysis. These transformations include a criterion that allow us to control the reconstruction process, and they can avoid some of the undesirable effects shown by the classical openings and closings by reconstruction in some situations. In particular, we can limit the excessive flat zones reconstruction that sometimes arises. In addition, we can separate, if desired, certain regions in images. We use these two effects in image segmentation. Among others we use morphological techniques to segment an input image, and we take advantage of the desirable properties of those filters to compute markers or significant regions. We will focus in this work especially on medical images applications.