Classification of Audio Signals Using a Bhattacharyya Kernel-Based Centroid Neural Network

A novel approach for the classification of audio signals using a Bhattacharyya Kernel-based Centroid Neural Network (BK-CNN) is proposed and presented in this paper. The proposed classifier is based on Centroid Neural Network (CNN) and also exploits advantages of the kernel method for mapping input data into a higher dimensional feature space. Furthermore, since the feature vectors of audio signals are modelled by Gaussian Probability Density Function (GPDF), the classification procedure is performed by considering Bhattacharyya distance as the distance measure of the proposed classifier. Experiments and results on various audio data sets demonstrate that the proposed classification scheme based on BK-CNN outperforms conventional algorithms including Self-Organizing Map(SOM) and CNN.