Title Page
ABSTRACT
국문 초록
Contents
NOMENCLATURE 10
CHAPTER 1. INTRODUCTION 11
CHAPTER 2. RELATED WORK 15
2.1. Vision-Language Models 15
2.2. Data Augmentation 15
2.3. Prompt-Learning 17
2.4. Few-Shot Learning (FSL) 17
CHAPTER 3. METHODOLOGY 19
3.1. CLIP Review 19
3.2. CoOp and CoCoOp Review 20
3.3. Mixup Data Augmentation Review 22
CHAPTER 4. EXPERIMENTS 24
4.1. Mixup Data Augmentation 25
4.2. Random Mixup Data Augmentation 30
CHAPTER 5. CONCLUSION 36
Bibliography 37
Table 1. Comparison of CoOp with and without Mixup Data Augmentation performances in base-to-new generalization environment setting with 16 shots... 28
Table 2. Comparison of CoCoOp with and without Mixup Data Augmentation performances in base-to-new generalization environment setting with 16 shots... 29
Table 3. Comparison of CoOp with and without Random Mixup Data Augmentation performances in base-to-new generalization environment setting... 32
Table 4. Comparison of CoCoOp with and without Random Mixup Data Augmentation performances in base-to-new generalization environment setting... 34
Figure 1. Pseudo-code of Mixup Data Augmentation in CoOp and CoCoOp 26
Figure 2. Pseudocode of Random Mixup Data Augmentation 31