Abstract: In this work, a new approach for parallel-hierarchical(PH) networks learning having applied to the real-timeimage sequences in extended laser paths is proposed. Itis possible to synthesize PH network with learningabilities by using the general idea of artificialneural networks structured organization on the scheme:input layer - hidden layer - output layer. The 1stnetwork level should be used as input layer, nextlevels should be used as a hidden layer and the lastlevel should be used as an output one, as it istraditionally in artificial neural networks, Using themain PH network feature which determine the length ofnetwork algorithm it is possible to determine a numberof hidden layer elements. And in this way it formalizesthe procedure of obtaining the number of hidden layerelements. !4
展开▼