In this paper, we propose an optimal and automatic thresholding segmen
tation method of grey level images based on c-means fuzzy Kohonen clus
tering network. This method is very efficient for processing images wh
ere grey levels vary slowly between areas. The aim is to obtain the mo
st important and precisely information about contours outlines. The me
thod yields an optimization problem to minimize in relation to a vecto
r of (unknown) clusters centres.