Dissertation > Excellent graduate degree dissertation topics show

Algorithm Research on Blind Separation of Acoustic Mixtures

Author: JiaFan
Tutor: LiuFang
School: Xi'an University of Electronic Science and Technology
Course: Applied Computer Technology
Keywords: Blind Source Separation (BSS) Independent component analysis (ICA) Voice signal aliasing Multichannel Blind Deconvolution Immune Algorithm
CLC: TN912.3
Type: Master's thesis
Year: 2004
Downloads: 442
Quote: 4
Read: Download Dissertation

Abstract


Blind signal separation signal processing technology in recent years a hot research topic , due to its mixed-signal from the observed can recover the original signal, while the original signal and hybrid systems requires little prior knowledge , so the wireless communications , medical signal processing , image enhancement , and voice separation has a wide range of applications . Mixed speech signal separation is the original intention of the technology research , in the field of signal processing is a difficult problem . This paper describes the basic theory of blind source separation knowledge, research a variety of mixed- model speech separation algorithm and the actual environment mainly speech separation problems . The main work is as follows : 1. Discusses the frequency-domain blind separation algorithm actual length of the reverb environment Fast Fourier Transform (FFT) the relationship between the length , breaking the previous algorithm reverb environment requires FFT length greater than the length of the impulse response reverberation restrictions , studies show that the greater the length of the FIR filter is not obtained better separation effect . 2 . This article attempts to point a new multi-channel ICA algorithm for separating mixed voice signal is given based on FIR matrix algebra techniques new algorithm for time-frequency domain , experiments show that the algorithm is effective and fast convergence rate . 3 . BSS optimization algorithm based on the traditional algorithms, the actual circumstances of the separation of the speech signal collected , due to the presence of echoes , often leaving the poor performance of its global convergence separation performance degradation, global convergence article will better immune algorithm blind separation Algorithm for voice optimization phase , obtained good separation effect.

Related Dissertations

  1. Research of Routing Algorithm Based on Immune Algorithm for Wireless Sensor Networks,TN929.5
  2. A Study on the Electrical Stimulation System of Evoked EEG Signals and the Associated Signal Processing Algorithms,TN911.7
  3. Immune ant colony algorithm and its application,TP301.6
  4. Research on VRP Problem Based on Immune Genetic Algorithm,F252
  5. Research of Power System Reactive Power Optimization Based on Immune Ant Colony Algorithm,TP18
  6. Parameter Optimization for Support Vector Regression Based on Multiple Intelligent Computation Methods,TP18
  7. Research on Evolutionary Equation and Other Algorithms of Particle Swarm Optimization,TP18
  8. Community Intelligence Algorithm for Engineering Structure Optimization,TU311.41
  9. Constrained multi-objective immune algorithm based task scheduling strategy P2P,TP18
  10. The Improvement Clone Selection Algorithms and Their Application,TP18
  11. Fine-grained model based on GPU-accelerated parallel immune algorithm,TP18
  12. Artificial immune algorithm research and analysis in the robot pose and error compensation of,TP242.6
  13. Optimization Solution of Vehicle Routine Problem and Its Algorithm Design,O224
  14. Research on Hybrid Evolutionary Algorithm to Solve the Vehicle Routing Optimization Problem,TP301.6
  15. The Research of the Simulation with the Excitation System of the Synchronous Generator Based on TMS320F2812,TM31
  16. Artificial Immune Algorithm and Its Implementation on SOPC,TP18
  17. Immune Algorithms for Channel Assignment and Multi-user Detection of Cellular Network,TN929.5
  18. Research on Simulation of Imaging Process Based on Virtual CT System,TP391.41
  19. Medical Data Mining Application Based on Intelligent Algorithms,TP311.13
  20. Based on Improved Particle Swarm Optimization for Study of Selection of Distribution Center,TP301.6
  21. Immune Mobile-Agent - based IDS detector model,TP393.08

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