The objective of semantic segmentation inmicroscopic images is to extract the cellular, nuclear or tissuecomponents. This problem is challenging due to the largevariations of these components features (size, shape,orientation or texture). In this paper we improve thetechnique presented in [17] used to identify the epithelialnuclei (crypt) against interstitial nuclei in microscopic imagestaken from colon tissues. In the proposed enhanced approach,the crypt inner boundary is detected using the closingmorphological pyramid instead of morphological hierarchy.The outer crypt border is determined by the epithelial nuclei,overlapped by the maximal isoline of the inner boundary. Theuse of sampling in building the pyramid offers computationalefficiency, reduces the amount of used memory, increase therobustness and preserve the quality results. An analysis of thetwo approaches is performed considering the number of pixelsprocessed to create each level. Also the relation between thelevels of the hierarchical structures is established.


Smochina, C., Manta, V., & Kropatsch, W. (2011). Sampling Step Importance in Hierarchical Semantic Segmentation of Microscopic Images. In 15th International Conference on System Theory, Control and Computing 2011 (pp. 1–6). http://hdl.handle.net/20.500.12708/53911