Method Of Evaluating Leather Handle Character Based On Neural Network

Leather handle character was evaluated and classified using artificial neural network (ANN). Experimental mechanics parameters correlative to leather elasticity, softness, fullness and extendibility were obtained. The feedforward neural network used provided accurate model between the input parameters (bend constant (K-b), compression index (K-c), lastometer coefficient (K-l), attenuation constant (a(b)) and bend stress and attenuation speed constant (K)) and the output parameter (leather handle character grade). Although a few numbers of experimental data were used in revising the neural network, the results obtained are comparable to the traditional method of leather handle character evaluation.