“…Neural networks are known for their high processing speed, high classification accuracy, low sensitivity to noise,and easy thresholding capability to yield a binary image for automated detection. Recently they have gained a lot of attention and have been investigated in conjonction with several NDT techniques, the training being done either with experimental ( real or simulated defects), or with theoretical data [3][4]. However results obtained with training on experimental data do not seem as encouraging as with those obtained with theoretical one's for two major reasons, firstly a large training set is required which is difficult to obtain with real cases, secondly simulated cases are not representative enough.…”