语音信号处理-端点检测.pdfVIP

Third International Symposium on Intelligent Information Technology and Security Informatics A noise robust endpoint detection algorithm for whispered speech based on Empirical Mode Decomposition and entropy Xue-Dan Tan Ji-Hua Gu Dept. of Phys. Sci. and Tech. Dept. of Phys. Sci. and Tech Soochow University Soochow University Suzhou, China Suzhou, China tanxuedan@126.com Zhi Tao He-Ming Zhao Dept. of Phys. Sci. and Tech Dept. of Electron Soochow University Soochow University Suzhou, China Suzhou, China taoz@ Abstract—This paper proposes a novel endpoint detection algorithm identify whispered speech segments accurately. Both [6] and [7] to improve the speech detection performance in noisy environments. are well suited for endpoint detection in stationary noise. In the proposed algorithm, Empirical Mode Decomposition is However, most of noises are non-stationary. Each type of noise introduced to improve the perform