| 基于神经网络的病态嗓音复杂度识别 |
| |
| 引用本文： | 龚英姬.基于神经网络的病态嗓音复杂度识别[J].电声技术,2014(3):56-58. |
| |
| 作者姓名： | 龚英姬 |
| |
| 作者单位： | 河池学院物理与机电工程学院,广西宜州546300 |
| |
| 基金项目： | 河池学院青年课题资助项目（2012B-N002） |
| |
| 摘 要： | 基于嗓音发声系统的复杂性和病态嗓音高频端噪声特性明显,对正常、病态嗓音信号进行小波分解、重构,然后求重构的各频段信号的Lempel-Zvi复杂度,利用神经网络识别对各频段信号的Lempel-Zvi复杂度进行识别。实验结果表明:在高频段,病态嗓音复杂度识别率为84%,相对高于其他较低频段。通过模式识别的方法揭示了嗓音发声系统的病变时的噪声特性和复杂性。
|
| 关 键 词： | 病态嗓音 Lempel-Zvi复杂度 小波分析 神经网络 |
Identification of Pathological Voice Complexity Based on Neural Network |
| |
| GONG Yingji.Identification of Pathological Voice Complexity Based on Neural Network[J].Audio Engineering,2014(3):56-58. |
| |
| Authors: | GONG Yingji |
| |
| Affiliation: | GONG Yingji ( Department of Physics and Mechanical ＆ Electronic Engineering, Hechi University, Yizhou Guangxi 546300, China) |
| |
| Abstract: | Based on complexity of voice system and pathological voice＇ s noise character in high -frequency section ,firstly, wavelet decomposition, reconstruction for normal, pathological voice signals is used; secsend compute Lempel- Zvi complexity for each band signal that reconstructed ; then neural networks are used to identificate Lempel- Zvi complexity of each band signal. The result shows that, in high-frequency section, The complexity of pathological voice recognition rate is 84%, relatively higher than that of other low frequency. The pattern recognition method reveals the noise characteristics and complexity of voice system lesions. |
| |
| Keywords: | pathological voice Lempel- Zvi complexity wavelet analysis neural networks |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
