Dissertation > Excellent graduate degree dissertation topics show

The Study of Cough Signal Recognition Based on HMM-ANN Hybrid Model

Author: ZhengXiaoPing
Tutor: ShiRui
School: Chongqing University
Course: Computer Software and Theory
Keywords: Cough sound recognition HMM - ANN hybrid model MFCC with noise suppression technology
CLC: TN912.34
Type: Master's thesis
Year: 2011
Downloads: 53
Quote: 1
Read: Download Dissertation

Abstract


Cough as the most common symptoms of respiratory diseases, its frequency, intensity, type, duration and other parameters for clinical provide important information. Stage in the assessment of cough is usually just under the patient's chief complaint, the lack of objective measurement and quantitative assessment of the standard and analysis system. With the wide application of speech recognition and artificial intelligence, people eager to cough sound analysis and evaluation can achieve true human-computer interaction, so that the machine can understand like a human cough sound signal, and detection identified cough sound complete further research and analysis. Reference to domestic and international voice recognition technology and coughing sound Research HMM-ANN hybrid model is applied to cough sound recognition, hidden Markov model (HMM) and artificial neural network (ANN), and in the MATLAB platform simulation experiments. The main content of this paper are: 1 In this paper, on the basis of the analysis of cough sound generation mechanism and various types of properties on the acoustic preprocessing cough sound. Combined with the characteristics of the cough sound, throughout the pre-treatment process of sampling, filtering, pre-emphasis, sub-frame windowing, endpoint detection step study. 2 This paper analyzes the linear prediction coefficients, linear predictive cepstral coefficients (LPCC) and Mel Frequency Cepstral Coefficients (MFCC). Found by experimental comparison, based on the characteristics of the human ear MFCC better than cough speech recognition based on the channel model LPCC. Order to better reflect the dynamic characteristics of the cough sound, the noise suppression of the cough sound impact on MFCC second feature extraction, and RASTA MFCC combined to improve the conversion of the logarithmic function, i.e. noise suppression with the standard MFCC The first-order differential parameters as cough sound characteristics. Simulation results show that, relative to the other three parameters, with noise suppression technology the MFCC-order differential parameters improved for cough sound signal recognition effect. 3 using HMM the better timing modeling capabilities and ANN classification, the cumulative probability of all states in the upcoming HMM Viterbi decoding as a neural network input, the final results of the neural network nonlinear mapping output, create a coughing sound HMM -ANN hybrid model to study the cough sound learning and recognition algorithm based on the hybrid model. Based on the above, the entire cough speech recognition process in the MATLAB platform simulation results show that the cough mixture model-based speech recognition performance be improved to some extent.

Related Dissertations

  1. Multiple ANN/HMM Hybrid Used in Speech Recognition,TN912.34
  2. Research on Hmm-based Speech Recognition System of the Robot,TN912.34
  3. Research on Speaker Recognition System under the Visual C++ 6.0,TN912.34
  4. Study on Hybrid Model of Speech Recognition Based on HMM and PNN,TN912.34
  5. Research on the Key Technologies fo Speech Recognition for Robot Communication,TN912.34
  6. Speech Emotion Recogonition Research,TN912.34
  7. The Research and Implementation of Algorithm of Isolated Word Speech Recognition,TN912.34
  8. EMD-based speaker recognition,TN912.34
  9. Research on Chinese Continuous Speech Tone and Digit String Recognition System,TN912.34
  10. Application of MMCE Algorithm to Factor Analyzed Probability Statistic Models,TN912.34
  11. A Study of Feature Extraction Algorithm of Speech Recognition Confidence Measure,TN912.34
  12. Research of Speech Recognition of Isolated Word Under Embedded Linux Operation,TN912.34
  13. Design and Implementation of Voice Screening System for Public Security Based on Speaker Recognition,TN912.34
  14. RBF neural network optimized for speech recognition research,TN912.34
  15. FPGA-based speech recognition system design and implementation,TN912.34
  16. Research on Mandarin Connected Digit Speech Recognition System Based on HTK,TN912.34
  17. Conjunction speech recognition system codebook design and modeling hardware identification module,TN912.34
  18. The Design of Analog to Digital Conversion Circuit in Speech Recognition System,TN912.34
  19. Research and Realization of Embedded Speaker Independent Continuous English Speech Recognition Based on HMM,TN912.34
  20. The Research and Implement of Isolated Word Speech Recognition Algorithm Based on DTW Model,TN912.34

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