ABSTRACT | 第5-6页 |
摘要 | 第7-18页 |
List of Abbreviations | 第18-20页 |
缩略语对照表 | 第20-29页 |
Chapter 1 Introduction | 第29-43页 |
1.1 Research Backgrounds and Significances | 第29-33页 |
1.2 Research Status at Home and Abroad | 第33-37页 |
1.2.1 Researches on Generative Model | 第34页 |
1.2.2 Researches on Discriminative Model | 第34-36页 |
1.2.3 Researches on Sparse Coding based Classifier | 第36-37页 |
1.3 Research Motivations and Purposes | 第37-40页 |
1.4 Innovations and Structure Arrangements | 第40-43页 |
1.4.1 Innovations | 第40-41页 |
1.4.2 Structure Arrangement | 第41-43页 |
Chapter 2 Mathematical Foundation of Sparse Learning Machine | 第43-53页 |
2.1 Notations | 第43-44页 |
2.2 The Method of Constructing Sparse Learning Machine | 第44-47页 |
2.3 Generalization Bound of Sparse Learning Machine | 第47-52页 |
2.4 Conclusion | 第52-53页 |
Chapter 3 Sparse LS-SVM via Coupled Compressive Pruning | 第53-69页 |
3.1 Sparse LS-SVM via Coupled Compressive Pruning | 第53-56页 |
3.1.1 Compressive Sampling | 第53-54页 |
3.1.2 Coupled Compressive Pruning for Sparse LS-SVM | 第54-56页 |
3.2 Experimental Results and Analysis | 第56-67页 |
3.2.1 Benchmark Datasets | 第57-58页 |
3.2.2 Experiment 1:Performance of CP-LSSVM Using Different Recov-ery Algorithms on Sinc Data | 第58-61页 |
3.2.3 Experiment 2: Performance of CP-LSSVM under Different Sparsityon Sinc and Two-Spiral Data | 第61-62页 |
3.2.4 Experiment 3:Performance of CCP-LSSVM with the Variation ofMeasurement Matrix Size on Sinc Data | 第62-64页 |
3.2.5 Experiment 4: Comparison of Our Method with Other Related Prun-ing Methods on UCI Data | 第64-66页 |
3.2.6 Experiment 5:Test on the Robustness of CP-LSSVM and CCP-LSSVM | 第66-67页 |
3.3 Conclusion | 第67-69页 |
Chapter 4 Sparse LS-SVM for HIC | 第69-85页 |
4.1 Sparse LS-SVM | 第69-73页 |
4.1.1 Sparse LS-SVM via MMV | 第69-72页 |
4.1.2 Sparse LS-SVM via Coupled Compressed Sensing | 第72-73页 |
4.2 Experimental Results and Analysis | 第73-83页 |
4.2.1 Benchmark Hyperspectral Imageries | 第74-76页 |
4.2.2 Experiment 1:Comparison with SS-SVM and SS-LSSVM | 第76-78页 |
4.2.3 Experiment 2: Comparison of Different Spatial Information Acqui-sition Methods | 第78-81页 |
4.2.4 Experiment 3:Performance of CCS4-LSSVM with the Variation ofσ and μ | 第81-83页 |
4.3 Conclusion | 第83-85页 |
Chapter 5 Sparse LapSVM for HIC | 第85-103页 |
5.1 Sparse LapSVM for HIC | 第85-93页 |
5.1.1 LapSVM for HIC | 第85-88页 |
5.1.2 Kernel Propagation | 第88-91页 |
5.1.3 Compressed Sensing based Sparse LapSVM | 第91-93页 |
5.2 Experimental Results and Discussions | 第93-100页 |
5.2.1 Experiment 1:Investigation of the Algorithmic Performances | 第93-96页 |
5.2.2 Experiment 2: Performance of SS-LapSVM for Different Numberof Labeled Pixels | 第96-98页 |
5.2.3 Experiment 3:Performance of S3LapSVM-KP with the Variation ofWindowwidth and μ | 第98-100页 |
5.2.4 Experiment 4: Performance of S3LapSVM-KP with the Variation ofSparsity | 第100页 |
5.3 Conclusion | 第100-103页 |
Chapter 6 SCC-SNT for HIC | 第103-119页 |
6.1 SCC-SNT for HIC | 第103-109页 |
6.1.1 Tensor Algebra | 第103-104页 |
6.1.2 Related Sparse Coding based Classifier | 第104-106页 |
6.1.3 Sparse Coding based Classifier with Spatial Neighbor Tensor | 第106-109页 |
6.2 Experimental Results and Analysis | 第109-113页 |
6.2.1 Experiment 1:Investigation of the Algorithmic Performances | 第109-111页 |
6.2.2 Experiment 2: Investigation of the Number of Training Pixels | 第111-112页 |
6.2.3 Experiment 3:Investigation of the Compressive Sampling Ratios 846.2.4 Experiment 4: Investigation of the Parameters T and S | 第112-113页 |
6.3 Conclusion | 第113-119页 |
Chapter 7 HPSCC-SNT for HIC | 第119-133页 |
7.1 HPSCC-SNT | 第119-122页 |
7.2 Experimental Results and Analysis | 第122-131页 |
7.2.1 Investigation of the Algorithmic Performances | 第122-128页 |
7.2.2 Performance of HPSCC-SNT with Different Number of TrainingPixels | 第128-129页 |
7.2.3 Performance of HPSCC-SNT with the Variation of Parameter T | 第129-131页 |
7.3 Conclusion | 第131-133页 |
Chapter 8 Summarization and Prospect | 第133-135页 |
8.1 Summarization | 第133-134页 |
8.2 Prospect | 第134-135页 |
Bibliography | 第135-143页 |
Acknowledgements | 第143-145页 |
致谢 | 第145-146页 |
Resume | 第146-148页 |