Neural Network based Classification for Speaker Identification

Speaker Recognition is a challenging task and is widely used in many speech aided applications. This study proposes a new Neural Network (NN) model for identifying the speaker, based on the acoustic features of a given speech sample extracted by applying wavelet transform on raw signals. Wrapper based feature selection applies dimensionality reduction by kernel PCA and ranking by Info gain. Only top ranked features are selected and used for neural network classifier. The proposed neural network classifier is trained to assign a speaker name as label to the test voice data. Multi-Layer Perceptron (MLP) is implemented for classification, and the performance is compared with the proposed NN model.