Performance Improvement of RBF Network Using ART2 Algorithm and Fuzzy Logic System

Recognition of Identifiers from Shipping Container Images Using Fuzzy Binarization and Enhanced Fuzzy RBF Network

“…On the other hand, a small value may allow the classification of the input pattern into an existing cluster in spite of a considerable mismatch. Moreover, because many applications of image recognition based on the fuzzy ART network assign experiential values to the vigilance parameter and the learning rate used in the adjustment of connection weights, the success rate of recognition may deteriorate [9,10]. To correct this defect, we proposed an enhanced fuzzy ART network which adaptively adjusts the vigilance parameter and the learning rate in the learning process, and applied it to the middle layer in a fuzzy RBF network.…”

“…In order to evaluate the recognition performance of the Fig. 7 Learning and recognition algorithm of the enhanced fuzzy RBF network enhanced fuzzy RBF network, we compared the experiment results with those obtained using the conventional fuzzy RBF network and the ART2-based RBF network [9,13].…”

“…In the traditional fuzzy ART network, the vigilance parameter determines the allowable degree of mismatch between any input pattern and the stored patterns [8,9]. The vigilance parameter is the inverse of degree of tolerance.…”