Performance Comparison of Noise Classification Using Intelligent Networks

The performance of speech-processing systems such as speech coding, speech recognition has been degraded by background environmental noises such as car, bus, babble, factory, helicopter and street noise. Noise classification is essential to enhance the performance. A major step in the design of a signal classification system is the selection of a good set of features that are capable of separating the signals in the feature space. In order to reduce the effect of environmental noises on speech processing tasks, noise classification is required. In general, classification of noise module achieves an improvement in the performance of a system operating in the presence of background noise by dynamically adapting the processing algorithms to the particular type of environmental noise. In this proposed work fuzzy ARTMAP network and modified fuzzy ARTMAP network are used for classification of background noise signals. Further these results are compared with back propagation networks and with Radial Basis Function Network(RBFN). Our empirical results show that the fuzzy networks have robust features in distinguishing the different classes of noises.