Dissertation > Excellent graduate degree dissertation topics show

Feature Extraction and Recognition of EEG Based on the AR Model

Author: ZouQing
Tutor: TangJingTian
School: Central South University
Course: Biomedical Engineering
Keywords: BCI EEG AAR algorithm MVAAR algorithm Tasks Classification
CLC: R318
Type: Master's thesis
Year: 2008
Downloads: 405
Quote: 9
Read: Download Dissertation

Abstract


Brain - machine interfaces become a hot research topic in recent years, brain science , rehabilitation engineering , automatic control , the military field and biomedical engineering field because of its broad application value and prospect . The processing of the EEG is the emphasis and difficulty in the course of the study . Desynchronization EEG events / phase synchronization phenomenon as characteristic information , in-depth discussion of the adaptive algorithm based on AR model (AAR) and multi - variable parameters the the AAR model algorithm (MVAAR) EEG characteristics Extraction application . Describes a variety of model coefficients estimation method , using the Kalman filter method and fast QR decomposition respectively the AAR, MVAAR model coefficient estimates , appeared to maximize the characteristic information in the EEG signal . Linear analysis , task identification , classification and leave-one- three classifier based on the Mahalanobis distance . The introduction of mutual information , and kappa value , the concept of the value of the area in the ROC under the curve for the performance evaluation of the classification results . From the experimental results , MVAAR algorithm than AAR algorithms achieve a higher classification accuracy . AAR model to describe the non- stationary random characteristics of the EEG signal MVAAR algorithm to identify law subjectivity smaller low order generally select consistent high data simulation , multi- lead data input have a stronger versatility. Traditional linear classification , based on the Mahalanobis distance secondary classification , leave-one classification to achieve good results , but also have their own advantages and disadvantages . LDA and MDA algorithms are only determined by the mean and covariance of the data , when the larger of the two types of covariance matrix difference LDA method will exhibit a large deviation MDA approach will exhibit better results. Leave-one principle is simple , easy to implement , but when a large experimental data , the amount of computation and computing time will be the question we must consider . Different object because a different individual differences and test feedback period , its use classification of the same set of algorithms effect .

Related Dissertations

  1. A Study on the Coma Scales and Electroencephalogram for the Prediction of the Comatose Patient,R741.044
  2. Research of the Mental Fatigue Classification Based on HHT and Kernel Method,TN911.7
  3. EEG -based emotion recognition,TP391.4
  4. Neural electrophysiological dress Contact Comfort Research,Q42
  5. Machine learning method based automatic classification of EEG,TP181
  6. Research and design of the television remote control system based on brain-computer interface technology,TP872
  7. Study on Assessment Method of Mental Fatigue Based on Brain-Computer Interface,R318
  8. Intraoperative Electrocorticography Monitoring of Focal Excision of Intractable Epilepsy by Microsurgical Treatment,R742.1
  9. Evaluation of Sample Entropy and Bispectral Index on Monitoring Sedation Depth during Target-controlled Infusion of Propofol Anesthesia,R614
  10. EEG based on compressed sensing compressive sampling,TN911.7
  11. The Mental Arithmetic Study of EEG by Factor Analysis,R741.044
  12. EEG Analysis by Using High-Performance Computation,R318.0
  13. EEG Early-warning Indicators of Exploratory Study on Patients with Mild Cognitive Impairment or Alzheimer’ Disease,R749.16
  14. Quantitative Eeg Power Spectral Characteristics of Mild Cognitive Impairment,R749.1
  15. The EEG Experimental Study on the Function of Simple Self-relaxation-hypnosis in Releasing Learning Fatigue,B842.3
  16. The EEG Signal Collecting and Processing System of White Rat Based in PSoC and LabVIEW,TP274.2
  17. The Control Method Research for Prosthetic Hand Based on EEG,TP391.4
  18. Research on Analysis & Recognition Method for Multi-motor ERD/ERS Signal in BCI System,TN911.6
  19. Study on Extraction of α Rhythm in EEG Based on Wavelet Package and ICA,TN911.7
  20. Research and Design of the Sistem That Getting and Handling Mice Brain Signal,TH789
  21. Primary Study on the Rehabilitation Mechanism of Internet Addicction Disorder Based on ET Detecting Technology,R749.99

CLC: > Medicine, health > Basic Medical > Medical science in general > Biomedical Engineering
© 2012 www.DissertationTopic.Net  Mobile