Contents | 第1-7页 |
Figure Index | 第7-8页 |
Table Index | 第8-9页 |
Abstract | 第9-11页 |
摘要 | 第11-12页 |
1 Introduction | 第12-26页 |
·What Is Brain-Computer Interface | 第12-17页 |
·The Functional Parts of a BCI System | 第12-15页 |
·The State of Art BCI Systems | 第15-17页 |
·The Leading Groups in BCI Area | 第17-18页 |
·The Continuous Control in EEG-Based BCI and the Key Problems | 第18-21页 |
·An Outline of This Thesis and My Contributions | 第21-23页 |
·References | 第23-26页 |
2 Neurophysiological Background and Feature Extraction Methods in BCI | 第26-37页 |
·Neurophysiological Background | 第26-30页 |
·Sensorimotor Rhythms | 第26-27页 |
·Challenges | 第27-30页 |
·Feature Extraction Methods | 第30-36页 |
·Spatial Filters | 第31-32页 |
·Spectral Analysis Methods | 第32-34页 |
·Statistical Methods | 第34-36页 |
·References | 第36-37页 |
3 Bayesian Methods in BCI | 第37-53页 |
·The Basic Idea of Bayesian Methods | 第37-39页 |
·Generative and Discriminative Models | 第39-46页 |
·Fisher Linear Discriminant | 第40页 |
·Bayesian Logistic Model | 第40-41页 |
·Generative and Discriminative Gaussian Mixture Model | 第41-42页 |
·Hidden Markov Model | 第42-44页 |
·Support Vector Machine | 第44-46页 |
·Optimization Methods | 第46-51页 |
·Extended Baum-Welch Algorithm | 第46-48页 |
·Expectation-Maximization Method and Its Extension | 第48-49页 |
·Variational Bayesian Method | 第49-51页 |
·References | 第51-53页 |
4 Solving Unlabeled Problem Based on EM Algorithm | 第53-70页 |
·Motivation and Background | 第53-55页 |
·The Proposed Algorithm for Continuous Cursor Prediction | 第55-59页 |
·The Proposed Algorithm | 第56-59页 |
·Convergence Property | 第59页 |
·Implementation of the Proposed Algorithm | 第59-61页 |
·Feature Extraction | 第59-60页 |
·Classification Algorithm | 第60页 |
·Control and Decision | 第60-61页 |
·Experimental Results | 第61-67页 |
·Results and Comparisons | 第62-63页 |
·A Comparison of Error Rates with Different Initial Probabilities | 第63-66页 |
·A Further Study of the Efficacy of the Proposed Algorithm | 第66-67页 |
·Discussions | 第67-68页 |
·References | 第68-70页 |
5 GMM Based Accumulative Classifier | 第70-90页 |
·Motivation and Background | 第70-72页 |
·Accumulative Classification Method Based on Gaussian Mixture Model | 第72-79页 |
·Generative and Discriminative Gaussian Mixture Model | 第73-77页 |
·Combining the Outputs of GMM Classifier across Time | 第77-79页 |
·The Estimation of α_j | 第77-78页 |
·The Estimation of β_j | 第78-79页 |
·Experimental Results and Discussions | 第79-86页 |
·A Comparison of the Performance with Baseline Methods | 第81-83页 |
·A Further Study of the Efficacy of the Proposed Algorithm | 第83-86页 |
·Discussions | 第86-88页 |
·References | 第88-90页 |
6 Exploiting Information in BCI Data for Continuous Prediction | 第90-110页 |
·Motivation and Background | 第90-92页 |
·The ABLM Method for Continuous Prediction | 第92-99页 |
·The Proposed Method | 第93-97页 |
·The Variations of ABLM Method | 第97-99页 |
·Experimental Results | 第99-108页 |
·Results and Comparisons in the First Experiment | 第99-101页 |
·A Study of the Effect of the Initial Probability | 第101-103页 |
·Results and Comparisons in the Second Experiment | 第103-104页 |
·The Effect of Prior P(ω) | 第104-105页 |
·A Study of the Effect of Accumulative Power | 第105-107页 |
·Statistical Analysis of the Results | 第107-108页 |
·Discussions | 第108-109页 |
·References | 第109-110页 |
7 Combining Classifiers in BCI System | 第110-122页 |
·Why Combining Classifiers: Majority Voting VS Linear Combination | 第110-112页 |
·Accumulative Classification Method by Stacking | 第112-116页 |
·Bayesian Logistic Model | 第114-115页 |
·Combining Outputs of the Classifier | 第115-116页 |
·Experimental Results | 第116-119页 |
·A Comparison of the Performance with Baseline Methods | 第117-118页 |
·A Further Study of the Efficacy of the Proposed Algorithm | 第118-119页 |
·Discussions | 第119-120页 |
·References | 第120-122页 |
8 Conclusions and Future Works | 第122-124页 |
·Conclusions of My Work | 第122-123页 |
·Future Research Directions | 第123页 |
·References | 第123-124页 |
Appendix A: Proof of Equation (4.9) in Chapter Four | 第124-126页 |
Appendix B: Proof of Equations (6.7-6.10) in Chapter Six | 第126-130页 |
Acknowledgements | 第130-132页 |
Announcement | 第132-134页 |
Papers | 第134页 |