| 基于深层声学特征的端到端语音分离 |
| |
| 引用本文： | 李娟娟,王丹,李子晋. 基于深层声学特征的端到端语音分离[J]. 计算机系统应用, 2019, 28(10): 1-7 |
| |
| 作者姓名： | 李娟娟 王丹 李子晋 |
| |
| 作者单位： | 复旦大学 计算机科学技术学院, 上海 201203,盲信号处理国家级重点实验室, 上海 200434,中国音乐学院 音乐科技系, 北京 100101 |
| |
| 基金项目： | 国家自然科学基金（61671156）；北京市社会科学基金（17YTC028） |
| |
| 摘 要： | 提出基于深层声学特征的端到端单声道语音分离算法，传统声学特征提取方法需要经过傅里叶变换、离散余弦变换等操作，会造成语音能量损失以及长时间延迟.为了改善这些问题，提出了以语音信号的原始波形作为深度神经网络的输入，通过网络模型来学习语音信号的更深层次的声学特征，实现端到端的语音分离.客观评价实验说明，本文提出的分离算法不仅有效地提升了语音分离的性能，也减少了语音分离算法的时间延迟.
|
| 关 键 词： | 语音分离 声学特征 深度神经网络 语音原始波形 端到端模型 |
| 收稿时间： | 2019-03-12 |
| 修稿时间： | 2019-04-04 |
End-to-End Speech Separation Based on Deep Acoustic Feature |
| |
| LI Juan-Juan,WANG Dan and LI Zi-Jin. End-to-End Speech Separation Based on Deep Acoustic Feature[J]. Computer Systems& Applications, 2019, 28(10): 1-7 |
| |
| Authors: | LI Juan-Juan WANG Dan LI Zi-Jin |
| |
| Affiliation: | School of Computer Science, FudanUniversity, Shanghai 201203, China,National Key Laboratory of Blind Signal Processing, Shanghai 200434, China and Department of Music Technology, China Conservatory of Music, Beijing 100101, China |
| |
| Abstract: | An end-to-end single channel speech separation algorithm based on deep acoustic feature is proposed. The traditional acoustic feature extraction methods require the Fourier transform, discrete cosine transform and other operations. This will cause speech energy loss and long latency. In order to improve these problems, the original waveform of the speech signal is used as an input to a deep neural network, deeper acoustic features of the speech signal are learned through a network model. Objective evaluation shows that the proposed algorithm not only improves the performance of speech separation effectively, but also reduces the time delay of speech separation algorithm. |
| |
| Keywords: | speech separation acoustic feature deep neural network speech original waveform end-to-end model |
|
| 点击此处可从《计算机系统应用》浏览原始摘要信息 |
|
点击此处可从《计算机系统应用》下载全文 |
|
