声明
摘要
Abstract
List of Figures
List of Tables
Chapter 1 Introduction
1．1 Main Introduction
1．2 Research Objectives
1．3 Research Questions
1．4 Key Contributions
1．4 Thesis Organization
1．5 Chapter 1：Summary
Chapter 2 Related Work
2．1 Human Activity Recognition：ANN and DNN
2．2 Wearable Systems Types
2．3 Human Activity Recognition (HAIR) Limitations in a Literature review
2．4 Classification of Movement
2．5 Recognition Algorithms Components
2．6 Kinds of Features Investigated
2．7 Chapter 2：Summary
Chapter 3 Prayer Activity Recognition Model using CNN (PARC)
3．1 Process of Data Collection
3．1．1 Imbalanced Dataset
3．2 Pre-processing
3．3 Synthetic Minority Oversampling Technique (SMOTe)
3．4 Window Segmentation
3．5 Feature Extraction
3．6 Convolutional Neural Network Model for Prayer Activities
3．6．1 ConvolutionaI Neural Network Architecture
3．7 Algorithm：Convolutional Neural Network Model for Prayer Activities
3．8 Chapter 3：Summary
Chapter 4 Performance Evaluation，Results and Discussion
4．1 Performance Evaluation
4．2 Experimental Results and Analysis
4．2．1 Performance Evaluation at Left Hand Body Position：Imbalanced Dataset
4．2．2 Performance Evaluation at Left Hand Body Position：Balanced Dataset
4．2．3 Performance Evaluation at Right Hand Body Position：Imbalanced Dataset
4．2．4 Performance Evaluation at Right Hand Body Position：Balanced Dataset
4．2．5 Performance Evaluation at Trouser Pocket Body Position：Imbalanced Dataset
4．2．6 Performance Evaluation at Trouser Pocket Body Position：Balanced Dataset
4．2．7 Summarized Results Comparison of four Body Positions：Imbalanced Dataset
4．2．8 Summarized Results Comparison of four Body Positions：Balanced Dataset
4．2．9 Performance Metric Comparison of four Body Positions：Imbalanced Dataset
4．2．10 Performance Metric Comparison of four Body Positions：Balanced Dataset
4．2．11 Comparison with previous work and other classifiers
4．2．12 Chapter 4：Summary
Chapter 5 Conclusion
5．1 Conclusion
5．2 Future Direction
REFERENCES
Chapter 6 Publications List
山东大学;
