Some key factors in speaker recognition using neural networks approach

Some key factors influencing the performance of a speaker recognition system using a neural network approach are addressed. The selection of the input pattern affects the complexity and performance of the neural network. In the experiment, seven kinds of input patterns were selected. The complexity of the distribution of every kind of input pattern is different, which affects the training time and recognition rate. The network with the most suitable input pattern gives the best performance and requires the best training time. The optimal scale of the neural network is affected by the complexity of the problem. Applying a subnetwork approach may keep the scale of the network acceptable or reasonable as the number of speakers increases. A method for text-independent speaker recognition is also described