Title Page
Contents
Abstract 7
I. Introduction 8
II. Related work 9
2.1. Super-Resolution Basic Theoretical Knowledge 9
(1) Super-Resolution methods based on Interpolation 9
(2) Super-Resolution methods based on Reconstruction 15
(3) Super-Resolution method based on Learning 18
2.2. Super-Resolution Methods based on Deep Learning 21
2.3. Residual Map 23
III. Residual U-Net 25
3.1. U-Net for Super-Resolution 26
3.2. Dense Skip Connections 28
3.3. Overall Framework 31
3.4. Evaluation Indicators 32
3.5. Training 34
IV. Experiment 38
4.1. Dataset 38
4.2. Network Architecture 39
4.3. Experimental Implementation Details 41
4.4. Experimental Results and Analysis[원문불량;p.39] 42
V. Conclusion 48
VI. References 49