A new self-generating prototypes method based on SGNT is presented. This method uses reference patterns as initial
prototype. This procedure can be implemented in a SGNT with specific architecture consisting of one root and the initial
class number of reference patterns. The leaf in SGNT is defined with prototype vector, learning vector, center property
vector and distant property vector. After training, prototype set are outputted. The main advantage of this method is that
both the number of prototypes and their locations are learned from the training set without much human intervention.
Experiments with synthesis and real color image the excellent performance of this classification scheme as compared to
existing K-nearest neighbor (K-NN) and Learning vector quantization (LVQ) algorithm.
A new adaptive remote sensing image fusion classification based on the Dempster-Shafer theory of evidence is presented.
This method uses a limited number of prototypes as items of evidence, which is automatically generated by modified
Fuzzy Kohonen Clustering Network (FKCN). The class fuzzy membership of each prototype is also determined using
reference pattern set. For each input vector a basic probability assignment (BPA) function are computed based on these
distances and on the degree of membership of prototypes to each class. And lastly this evidence is combined using
Dempster's rule. This proposed method can be implemented in a modified FKCN with specific architecture consisting of
one input layer, a prototype layer, a BPA layer, a combination and output layer, and decision layer. The experimental
results show that the excellent performance of classification as compared to existing FKCN and basic DS fusion