Title :
Multi-scale kernels for short utterance speaker recognition
Author :
Wei-Qiang Zhang ; Junhong Zhao ; Wen-Lin Zhang ; Jia Liu
Author_Institution :
Dept. of Electron. Eng., Tsinghua Univ., Beijing, China
Abstract :
Short utterance is a great challenge for speaker recognition, for there is very limited data can be used for training and testing. To give a robust estimation, the amount of model parameters for the short utterance should be less than that for the long utterance; however, this may impede the models descriptive capability. In this paper, we propose a multi-scale kernel (MSK) approach to solve this problem. We construct a series of kernels with different scales, and combine them through multiple kernel learning (MKL) optimization. In this way, the robustness and scalability of the model will be both enhanced. The experimental results on NIST SRE 2010 10sec- 10sec dataset show that the proposed MSK method outperforms the traditional Gaussian mixture model supervector (GSV) followed by support vector machine (SVM) method.
Keywords :
estimation theory; learning (artificial intelligence); optimisation; speaker recognition; MKL optimization; multiscale kernel learning; robust estimation; short utterance speaker recognition; Kernel; NIST; Robustness; Speaker recognition; Support vector machines; Training; Vectors; multi-scale kernel; short utterance; speaker recognition;
Conference_Titel :
Chinese Spoken Language Processing (ISCSLP), 2014 9th International Symposium on
Conference_Location :
Singapore
DOI :
10.1109/ISCSLP.2014.6936594