Modified Fuzzy Kohonen Clustering Network for image segmentation
Nahla Ibraheem Jabbar, Syed Ismail Ahson · 2010
This study derives a new interpretation for Fuzzy Kohonen Neural Network in parameter m. Derivation of new algorithm is accomplished by adaptive m. The learning rate in each iteration is effected by updating m ascending or descending. This yields automatic control to distribution of learning rate in the neighborhood of all data nodes at each iteration. This experiment evaluates the modified FKCN algorithm in terms of ability of segmentation of different regions in the image. Both the analytical and experimental data reported indicates that the choice of m in the range.