Title : INFORMATION SYSTEMS BASED ON NEURAL NETWORK AND WAVELET

This paper suggests a novel methodology for building robust information processing systems based on wavelet and neural network methods to be used in decision making tasks when image information is involved as well as in prediction and modelling tasks when signal information should be processed. It is proposed that the efficiency of such systems is increased when they simultaneously process input information in both its original and wavelet transformed form. Artificial Neural Network (ANN) technology is invoked to fuse these two different types of input. The suggested novel approach is applied to the design of two novel information processing systems. Namely, a quality control decision making system as well as a signal prediction and modelling system are used to illustrate the proposed methodology. More specifically, the first one offers a solution to the problem of defect recognition from images, that can find applications in building robust quality control vision based systems. Such applications can be found in the production lines of textile, integrated circuits, machinery, etc. The second one attempts to improve the quality of time series prediction and signal modelling systems by dealing with the peak prediction problem appeared in such tasks. Its properties are exhibited in an NMR signal since many peaks are normally encountered in them. The increased accuracy obtained by our approach shows that it deserves the attention of the designers of effective information processing systems handling images and signals, targeted to decision making, modelling and prediction tasks.