Improvement on automatic speaker gender identification using classifier fusion

In this paper, a two layer classifier fusion technique is proposed for automatic gender identification (AGI). The first layer is an acoustic classification layer for mapping MFCC acoustic feature space to score space. In this layer, a divisive clustering is proposed for dividing the speakers from each gender to some classes of speakers having similar vocal articulatory. The second layer is a back-end classifier that receives the vectors of fused likelihood scores from the first layer. GMM, SVM and MLP classifiers were evaluated in the middle and back-end layers. 96.53% gender classification accuracy was obtained on OGI multilingual corpus which is much better than the performance obtained by traditional AGI methods.