全局信息融合的汉语方言自动辨识-计算机工程与应用.PDFVIP

160 2017 ，53（17） Computer Engineering and Applications 计算机工程与应用 全局信息融合的汉语方言自动辨识 用 1 1 1 1 1 1 2 邱远航 ，顾明亮 ，马 勇 ，金 赟 ，韩 军 ，赵冬梅 ，赵呈昊 应 1 1 1 1 1 1 2 QIU Yuanhang , GU Mingliang , MA Yong , JIN Yun , HAN Jun , ZHAO Dongmei , ZHAO Chenghao 与 1.江苏师范大学 物理与电子工程学院，江苏 徐州 221116 程 2.江苏师范大学 电气工程及自动化学院，江苏 徐州 221116 g r 1. School of Physics and Electronic Engineering, Jiangsu Normal University, Xuzhou, Jiangsu 221116, China 工 o 2. School of Electrical Engineering Automation, Jiangsu Normal University, Xuzhou, Jiangsu 221116, China . 机 j a QIU Yuanhang, GU Mingliang, MA Yong, et al. Automatic identification of Chinese dialects based on global infor- 算 e mation fusion. Computer Engineering and Applications, 2017, 53 （17）：160-165. c 计 . Abstract: A new method of Chinese dialects identification based on Identity vector （I-vector ）combined with prosodic w information is proposed. The high-dimensional super-vector is mapped to a low-dimensional I-vector representation by w Total Variability （TV ）model. Channel compensation and feature dimension reduction are also performed. Chinese is a typ- w ical language with a tone and Chinese dialects have obvious differences among rhythm, stress and other rhythmic struc- ture. The serial fusion of I-vectors with global prosodic information can improve the distinguishability of