| 合成语音检测方法的研究现状及展望 |
| |
| 引用本文： | 魏为民,刘畅,才智,孟繁星. 合成语音检测方法的研究现状及展望[J]. 上海电力学院学报, 2022, 38(1): 75-81 |
| |
| 作者姓名： | 魏为民 刘畅 才智 孟繁星 |
| |
| 作者单位： | 上海电力大学 计算机科学与技术学院 |
| |
| 基金项目： | 上海市自然科学基金(16ZR1413100) |
| |
| 摘 要： | 随着语音诈骗事件的发生, 自动说话人验证系统已被广泛应用。传统的检测合成语音的方法包括利用频谱信息、倒谱系数特征以及动态声学特征等。随着机器学习的发展, 深度学习算法也被应用于合成语音检测且取得了不错的成果。介绍了几种检测方法的基本原理及部分研究成果, 并对以后的发展方向进行了展望。
|
| 关 键 词： | 合成语音检测 频谱信息 神经网络 |
| 收稿时间： | 2020-02-13 |
Research Status and Prospect of Synthetic Speech Detection |
| |
| WEI Weimin,LIU Chang,CAI Zhi,MENG Fanxing. Research Status and Prospect of Synthetic Speech Detection[J]. Journal of Shanghai University of Electric Power, 2022, 38(1): 75-81 |
| |
| Authors: | WEI Weimin LIU Chang CAI Zhi MENG Fanxing |
| |
| Affiliation: | School of Computer Science and Technology, Shanghai University of Electric Power, Shanghai 200090, China |
| |
| Abstract: | With the occurrence of voice fraud, automatic speaker verification system has been widely used.Traditional methods of detecting synthetic speech include using spectral information, Q cepstrum coefficient characteristics and dynamic acoustic characteristics.With the development of machine learning, deep learning algorithm has also been applied to synthetic speech detection and achieved good results.The basic principle and some research results of several detection methods are introduced, and the future development direction is prospected. |
| |
| Keywords: | synthetic speech detection spectral information neural network |
|
| 点击此处可从《上海电力学院学报》浏览原始摘要信息 |
|
点击此处可从《上海电力学院学报》下载全文 |
|
