Dissertation > Excellent graduate degree dissertation topics show
The Research of Voice Activity Detection Based on Characters in Noise Environment
Author: ZhaoLiXia
Tutor: ZhaoHuan
School: Hunan University
Course: Computer Science and Technology
Keywords: Voice activity detection Feature Entropy Support Vector Machine
CLC: TN912.3
Type: Master's thesis
Year: 2010
Downloads: 142
Quote: 3
Read: Download Dissertation
Abstract
|
The purpose of the speech endpoint detection is to determine the beginning and end of the speech , from the period of the signal containing speech voice signal processing front-end operations , is widely used in the field of speech enhancement , speech coding , speech recognition . Speech endpoint detection method feature-based and model-based two categories , the model-based approach is more complex , the ability to adapt to the environment , and feature-based approach is relatively simple and has some anti-noise capability , this method requires to find some kind of able to distinguish between voice and noise characteristics of robustness . In this paper, the speech endpoint detection method based on the characteristics of a study . Proposed a new distance-based entropy detection algorithm for the detection algorithm based on spectral entropy at low signal-to-noise ratio under the disadvantage of poor robustness . The algorithm uses the probability density of the robustness of the change of the entropy and cepstrum calculation method , a series of calculations of the pre-treated with a noise signal point cepstral coefficients , cepstral coefficients Euclidean distance , Euclidean distance construct the probability density function of the probability density function from the entropy feature distance entropy Finally, a dual - threshold distinction between speech and noise . This paper proposed a support vector machine - based multi-feature detection algorithm . Seeking SNR noisy signal detection algorithm based on support vector machine modified zero the rate and AMMM three characteristics , three characteristics to form a feature matrix , use some noisy signal of support vector machine training , using the training after support vector machine automatically distinguish between voice and noise . This article experiments with noise signal from the noise of France aurora2.0 library of clean voice and Noisex92 of noise library mixture , and simulation experiments using MATLAB tools , the experimental results show that , the two endpoint detection algorithm proposed in this paper has certain robustness , even at lower signal-to-noise ratio better distinguish between voice and noise .
|
Related Dissertations
- Research on Automatic Detection Algorithm for Substructure Distress of Highway Pavement Based on SVM,U418.6
- The Research of Steel’s Representation in Architecture,TU-0
- Numerical Simulation of Polarized Radiative Energy and Entropy Transfer,TK124
- Research of Irreversible Loss in Flow and Mexed Prosess,TK12
- Quantitative evaluation model based on information entropy of the classroom observation,G632.4
- ISAR Imaging Simulation of Space Targets and Target Recognition Based on ISAR Images,TN957.52
- Research on Autamatic Music Structrue Analysis,TN912.3
- The Research and Implemention of Image Retrieval Based on User Interested Feature,TP391.41
- Research on the Classification Based on the Reconstruction of Solder Joint,TP391.41
- Syntactic Features Based Pronoun Resolution,TP391.1
- The Software Component Modeling Method Based on Feature and Its Application in VMI Management System,TP311.52
- Research on Feature Extraction and Classification of Pulse Waveform for Cholecystitis and Nephrotic Syndrome Diagnosis,TP391.41
- Application of Q-Learning in the Content-Based Image Retrieval Technology,TP391.41
- Research on Domain Entity Attribute and Event Extraction Technology,TP391.1
- Research on Transductive Support Vector Machine and Its Application in Image Retrieval,TP391.41
- Research on Feature Extraction and Classification of Tongue Shape and Tooth-Marked Tongue in TCM Tongue Diagnosis,TP391.41
- Research on Extraction and Tracking of People’s Opinion,TP391.1
- Research on Text Classification Based on Biomimetic Pattern Recongnition,TP391.1
- Research on Removal Algorithm of Shadows in Image Segmentation,TP391.41
- Research on Visual Measurement for Spacecraft Rendezvous and Approach,TP391.41
- Research on the Image Real-Time Acquisition, Storage and Image Processing System,TP391.41
CLC: > Industrial Technology > Radio electronics, telecommunications technology > Communicate > Electro-acoustic technology and speech signal processing > Speech Signal Processing
© 2012 www.DissertationTopic.Net Mobile
|