参考文献及笔记

TODO

1. 编写基本的项目框架代码，实现一种方法（MetaNets），后面考虑扩展比较 AdaIN （Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization）和特征变换（Universal Style Transfer via Feature Transforms）
2. 学习评估指标 风格转移指标（Style Transfer Indicator）Evaluation of Painting Artistic Style Transfer Based on Generative Adversarial Network
3. 学习 PyTorch 框架(2)
   • 迁移学习
   • 实验跟踪
   • 论文复制
   • 模型部署

涉及文献

1. VGG——Very Deep Convolutional Networks for Large-Scale Image Recognition

2. ResNet——Deep Residual Learning for Image Recognition

3. U-Net: Convolutional Networks for Biomedical Image Segmentation

4. GAN——Generative Adversarial Networks

5. Conditional Generative Adversarial Nets

6. PatchGAN 起源 —— Image-to-Image Translation with Conditional Adversarial Networks

7. 将PatchGAN扩展为多尺度（3级金字塔）—— High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

8. Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks

9. Meta Networks for Neural Style Transfer

风格迁移实战代码参考

1. Image Style Transfer Using Convolutional Neural Networks：https://github.com/b06b01073/style-transfer/tree/main