文章目录
- 基本思路:Introdution
- 拉普拉斯金字塔
- LAPGAN
- LAPGAN的正向过程
- 训练过程
基本思路:Introdution
这篇论文的一个基本方法逻辑是:利用拉普拉斯金字塔结构,将拉普拉斯金字塔中的每一层利用CGAN结构来逐步生成一个高精度的图像。
In contrast, we propose an approach that is
able to generate plausible looking scenes at 32 × 32 and 64 × 64.
To do this, we exploit the multi-scale structure of natural images, building a series of generative models, each of which captures
image structure at a particular scale of a Laplacian pyramid。
At each scale we train a convolutional networkbased generative model using the Generative Adversarial Networks。