This paper presents a novel scheme for considering the frame-level speaker relevancy during i-vector extraction for speaker recognition. In the proposed system, the frame-level point-wise mutual information is utilized to directly modify the Baum-Welch statistics in order to extract a robust i-vector. Furthermore, a method for computing the frame-level speaker relevancy using deep neural network (DNN) analogous to the DNN used in robust automatic speech recognition (ASR) is proposed. The results show that the modified i-vectors obtained using the proposed methods outperformed the conventional i-vectors.
CITATION STYLE
Kang, W. H., Cho, W. I., Jang, S. Y., Lee, H. S., & Kim, N. S. (2017). I-vector extraction using speaker relevancy for short duration speaker recognition. In Lecture Notes in Electrical Engineering (Vol. 449, pp. 79–87). Springer Verlag. https://doi.org/10.1007/978-981-10-6451-7_10
Mendeley helps you to discover research relevant for your work.