Semantic Photo Manipulation with a Generative Image Prior
Semantic Photo Manipulation with a Generative Image Prior
复制标题
DOI：
10.1145/3306346.3323023
复制
发表时间：
2019-07-01
影响因子：
6.2
通讯作者：
Torralba, Antonio
中科院分区：
文献类型：
--
作者：
Bau, David;Strobelt, Hendrik;Torralba, Antonio
尽管 GAN 最近在根据用户草图、文本或语义标签等输入合成图像方面取得了成功，但使用 GAN 操纵现有自然照片的高级属性仍然具有挑战性，原因有二。首先，GAN 很难精确地再现输入图像。其次，经过处理后，新合成的像素通常不适合原始图像。在本文中，我们通过将 GAN 先前学习的图像适应单个图像的图像统计来解决这些问题。我们的方法可以准确地重建输入图像并合成新内容，与输入图像的外观一致。我们在几个语义图像编辑任务上演示了我们的交互式系统，包括合成与背景一致的新对象、删除不需要的对象以及更改对象的外观。与几种现有方法的定量和定性比较证明了我们方法的有效性。
Despite the recent success of GANs in synthesizing images conditioned on inputs such as a user sketch, text, or semantic labels, manipulating the high-level attributes of an existing natural photograph with GANs is challenging for two reasons. First, it is hard for GANs to precisely reproduce an input image. Second, after manipulation, the newly synthesized pixels often do not fit the original image. In this paper, we address these issues by adapting the image prior learned by GANs to image statistics of an individual image. Our method can accurately reconstruct the input image and synthesize new content, consistent with the appearance of the input image. We demonstrate our interactive system on several semantic image editing tasks, including synthesizing new objects consistent with background, removing unwanted objects, and changing the appearance of an object. Quantitative and qualitative comparisons against several existing methods demonstrate the effectiveness of our method.