Scoring methods for normalized kernels for multi-level speaker verification

In this article the text-independent speaker verification problem is considered. In the presented system each conversation side is represented as a vector lying on the unit hypersphere. These vectors are compared by an inner product which produces similarity scores. In this article classical score normalization methods (z-norm and t-norm) are analyzed and compared with the support vector machines (SVMs). Next, the simplified support vector machine algorithm is proposed with the advantage of speed. All presented methods are experimentally evaluated as a part of the multi-level speaker verification system.