论文推荐 | Cycle-GAN：通过双向循环实现一个超棒的图像风格转换器

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

（Cycle-GAN：通过双向循环实现一个超棒的图像风格转换器）

Zhu Jun-Yan /Park Taesung /Isola Phillip /Efros Alexei A.

推荐原因

图像到图像的转换，或者说图像风格转换，传统的想法是直接通过 U-Net 之类的结构做像素级别的一对一转换。显然这种做法不能允许图像的大幅度改动，以及还有一个更让人难受的问题是，一定需要内容对应、风格不同的图片对（比如同一位置同一视角仅时间不同的森林照片）作为训练数据才能训练。

GANs 这样的生成器-鉴别器的架构其实很有潜力完成图像转换任务，只不过并不是直接拿来就能用。首先要把 GANs 生成器的从随机种子生成图像换成从给定图像生成新图像；同时，为了保留图像中的内容信息，作者们设计了损失函数，要求转换后的图像可以再被转换回原图（形成循环，也就是CycleGAN中cycle的由来）。这种设计极大地提升了图像转换任务中的模型表现，而且可以不再需要对应的图片对，只需要是两个不同分布的图像就可以。

想法巧妙，数据要求低，执行效果好，CycleGAN也从而成为了2017年最热门的模型之一。

来自AI研习社用户@杨 晓凡的推荐

摘要

Image-to-image translation is a class of vision and graphics problems where the goal is to learn the mapping between an input image and an output image using a training set of aligned image pairs. However, for many tasks, paired training data will not be available. We present an approach for learning to translate an image from a source domain $X$ to a target domain $Y$ in the absence of paired examples. Our goal is to learn a mapping $G: X \rightarrow Y$ such that the distribution of images from $G(X)$ is indistinguishable from the distribution $Y$ using an adversarial loss. Because this mapping is highly under-constrained, we couple it with an inverse mapping $F: Y \rightarrow X$ and introduce a cycle consistency loss to push $F(G(X)) \approx X$ (and vice versa). Qualitative results are presented on several tasks where paired training data does not exist, including collection style transfer, object transfiguration, season transfer, photo enhancement, etc. Quantitative comparisons against several prior methods demonstrate the superiority of our approach.

论文查阅地址（扫描二维码直达）：

http://ai.yanxishe.com/page/paperDetail/26