Dissertation > Excellent graduate degree dissertation topics show

Research on Technology of Speaker Recognition Based on VQ and HMM

Author: FangZuoJun
Tutor: WangSheGuo
School: Hebei University of Engineering
Course: Applied Computer Technology
Keywords: Speaker Recognition Hidden Markov Models Vector quantization Linear Predictive Cepstral Mel Frequency Cepstral
CLC: TN912.34
Type: Master's thesis
Year: 2008
Downloads: 165
Quote: 0
Read: Download Dissertation

Abstract


Speaker recognition is based on the voice signal reflected the personality traits of the physiology and behavior of the speaker automatically identify the identity of the speaker , a biometric authentication technology can be widely applied to telephone banking , in areas such as database access, the computer remote login , secure authentication , automatic control speaker recognition with such a broad application prospects in the field of biometric authentication technology in recent years more and more people's attention. How to get the personality traits of the speaker and to choose reasonable speaker recognition method to improve the recognition rate and reliability of the speaker recognition system , the main contents of this paper . This article recognize the fundamental principle and system structure through the analysis of the speaker , to examine the existing speaker recognition technology , research based on linear prediction cepstral coefficients (Linear Prediction Cepstrum Coefficient, LPCC) and Mel-frequency cepstral coefficients ( MelFrequency Cepstrum Coefficient MFCC ) the combination of the second feature extraction methods , as well as vector quantization (Vector Quantization, VQ) and the hidden the horse Markov model (Hidden Markov Model, HMM) recognition method . The design of the system in this article is the text related to the speaker recognition system . In order to obtain an effective speech segments and better extraction of the characteristic parameters of the speaker , the first speaker's speech signal denoising, pre-emphasis , sub-frame, pretreatment such as windowing and endpoint detection , LPCC and MFCC feature weighting , characteristic difference , combinations of features and feature selection the secondary characteristic parameters , and in front of the HMM model using the VQ technical design codebook for each speaker , to avoid causing the error accumulation effect , last Baum-Welch algorithm and Viterbi algorithm for speaker training and recognition . Experiments show that the LPCC and MFCC secondary characteristic parameters , and VQ HMM model combined method of speaker recognition system has a high recognition rate and low error rate .

Related Dissertations

  1. Research on Lapped Transform and Vector Quantization Based Image Coding Algorithm and Applications,TN919.81
  2. Research on Quantum Particle Swarm Optimization and Its Application to Codebook Design of Image Vector Quantization,TP301.6
  3. Study on Compressed Sensing Algrithom and Its Application in Vector Quantization,TN911.7
  4. Image semantic annotation of blocks - a global feature extraction method,TP391.41
  5. Pre ~ 2VOD: one kind of VOD/P2P VCR operation support system,TN948.64
  6. Moving target trajectory based recognition system for human-computer interaction,TP391.41
  7. Research on Human Abnormal Behavior Detection and Recognition in Intelligent Video Surveillance,TP391.41
  8. Study on Artificial Bandwidth Extension of Speech Signal,TN912.3
  9. Researches on Hyperspectral Image Lossless Compression Algorithm Based on Vector Quantization,TP751.1
  10. Research on Very Low Bit Rate Speech Coding,TN912.3
  11. Research on Speech-Like Modulation and Demodeulation in End-to-End Secure Communication,TN918
  12. Research on Key Techniques of AMR-WB+ Algorithm and Coded Stream Recognition,TN912.3
  13. Design and Implementation of Speaker Recognition System Based on Windows CE,TN912.34
  14. Research of Low-bit-rate Speech Coding Algorithms Based on MELP,TN912.3
  15. The Research Based on the Text Irrelevant Speaker Distinguishes,TN912.34
  16. Highway traffic incident detection modeling and application research,U491.116
  17. Research on Efficient Parameter Quantization Algorithms for Low-bit-rate Speech Coding,TN912.3
  18. Audio Non-watermarking Based on the Audio Signal Feature,TP309.7
  19. Study on the Technology of Oil Pollution Class Based on Image Processing,TP274
  20. Based on SOM Intrusion Detection Algorithm of Research and Improvement,TP393.08
  21. Research of Image Compression Coding Based on Wavelet Transform and Vector Quantization,TP391.41

CLC: > Industrial Technology > Radio electronics, telecommunications technology > Communicate > Electro-acoustic technology and speech signal processing > Speech Signal Processing > Speech Recognition and equipment
© 2012 www.DissertationTopic.Net  Mobile