arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.10951cs.CVcs.AIcs.GR
收藏

绘画文字

Paint by Word

MIT CSAIL(麻省理工学院计算机科学与人工智能实验室);Harvard Graduate School of Design(哈佛大学设计研究生院)

Alex Andonian, Sabrina Osmany, Audrey Cui, YeonHwan Park, Ali Jahanian, Antonio Torralba, David Bau

摘要:

我们研究的问题zero-shot语义图像绘画。我们不是只使用具体的颜色或有限的语义概念集来绘制图像的修改,而是询问如何基于开放的全文描述创建语义绘制:我们的目标是能够指向合成图像中的一个位置,并应用任意的新概念,如“乡村”或“华丽”或“快乐的狗”。“为了做到这一点,我们的方法将最先进的现实图像生成模型与最先进的文本图像语义相似性网络相结合。我们发现,要进行大的更改,重要的是使用非梯度方法来探索潜在空间,并且重要的是放松GAN的计算以针对特定区域的更改。我们进行用户研究,将我们的方法与几个基线进行比较。

英文摘要:

We investigate the problem of zero-shot semantic image painting. Instead of painting modifications into an image using only concrete colors or a finite set of semantic concepts, we ask how to create semantic paint based on open full-text descriptions: our goal is to be able to point to a location in a synthesized image and apply an arbitrary new concept such as "rustic" or "opulent" or "happy dog." To do this, our method combines a state-of-the art generative model of realistic images with a state-of-the-art text-image semantic similarity network. We find that, to make large changes, it is important to use non-gradient methods to explore latent space, and it is important to relax the computations of the GAN to target changes to a specific region. We conduct user studies to compare our methods to several baselines.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑