An increasing amount of works has shown that the behavior exhibited by neural cells is a consequence not only of the biochemical processes in and outside the cells, but also of their respective morphology. For instance, a neuron presenting a complex dendritic tree will enhance the changes of establishing synaptic connections. The primary cells used in many morphological investigations are retinal ganglion cells, because their nearly planar dendritic arborization can be conveniently reduced to two dimensions. The extraction of the visual information is a complex process that involves several different phases. The objective of our study is to find the best combination of features that can properly classify these neurons inside each membership group with the smallest possible mistake.

Cat ganglion retinal cells.


In many pathological processes that involve megakaryopoiesis the size and shape of the cytoplasm and/or of the nucleus and the chromatin pattern can be extremely varied (multiple lobulated nuclei, coarse-grained nuclear chromatin, and so on). An unsupervised algorithm to help experts during the diagnosis phase has been developed: it distinguishes between pathological and normal cells. The experiment was conducted on high power magnification photomicrographs, taken from haematoxylin-and-eosin stained 3µ thick sections of formalin-fixed, paraffin embedded bone marrow trephine biopsies.

Segmentation of human megakaryocytes.