Neural network based fabric classification and blend composition analysis

Fabric classification plays an important role in the textile industry. Fabrics classification and determining the blend composition involves tedious work and time consumption. Knowledge based systems like artificial neural networks may be successfully employed to over come such problems. In this paper, a method is presented to classify and determine blend composition using neural networks. By inputting the mechanical properties measured from Kawabata evaluation system to the neural network, one can get the desired results. A comparative study in the results is made between the backpropagation and radial basis function algorithms to test the suitability of the same for the proposed textile application.