一种将图片梵高化（风格化）的神经网络_风闻

【来源龙腾网】

正文原创翻译：

​

In fine art, especially painting, humans have mastered the skill to create unique visual experiences through composing a complex interplay between the content and style of an image. Thus far the algorithmic basis of this process is unknown and there exists no artificial system with similar capabilities. However, in other key areas of visual perception such as obxt and face recognition near-human performance was recently demonstrated by a class of biologically inspired vision models called Deep Neural Networks. Here we introduce an artificial system based on a Deep Neural Network that creates artistic images of high perceptual quality. The system uses neural representations to separate and recombine content and style of arbitrary images, providing a neural algorithm for the creation of artistic images. Moreover, in light of the striking similarities between performance-optimised artificial neural networks and biological vision, our work offers a path forward to an algorithmic understanding of how humans create and perceive artistic imagery.

在美术领域，特别是绘画领域，人类已经掌握了通过合成图像的内容和风格之间的复杂相互作用创造独特的视觉体验，来创造独特视觉体验的技能。到目前为止，这一过程的算法基础尚不清楚（隐藏的神经网络层像个黑盒子），也不存在具有类似能力的人工系统。

然而，在视觉感知的其他关键领域，如物体和人脸识别，最近由一类被称为深度神经网络的、由生物启发的视觉模型展示了其接近人类的水平的性能。

在这里，我们介绍一个基于深度神经网络的人工系统，它可以生成高感知质量的艺术图像。

该系统通过使用神经表征分离和重组任意图像的内容和样式，为艺术图像的创建提供了一种神经算法。

此外，鉴于性能优化的人工神经网络和生物视觉之间惊人的相似之处，我们的工作为人类是如何创造和感知艺术意象的算法理解提供了一条前进的道路

Not so long ago we have been talking about deep neural networks,the technique that was inspired by the human visual system.

我久之前，我们才提到过深度神经网络，一项受人类视觉系统启发的机器学习技术。

It enables computers to learn things in a very similar way that the human would.

这项技术使得计算机以一种与人类非常相似的方式学习事物。

Neural networks are by no means perfect so do not worry,don’t quit your job, you’re good, but some applications are getting out of control.

仍需指出神经网络绝对是不完美的，所以别担心，它暂时不会让你丢掉工作，你很优秀，但有些神经网络的应用已经失控了。

In Google Deep Mind’s case is started to learn playing simple computer games and eventually showed a superhuman level plays in some cases.

在谷歌DeepＭind 的案例里，已经开始用神经网络来玩简单的电脑游戏，最后在某些情况下展示出了超越人类的能力水平。

I have run this piece of code and got some pretty sweet results that you can check out.

我也亲自运行过这段代码，结果很理想，你们可以自己去跑一跑。

So about this paper we have here today,what does this one do? You take photographs with your camera and you can assign it any painting and it will apply

this painting’s artistic style to it.

那么我们今天要脱水的这篇论文讲的神经网络是做什么的呢？拿相机照张相，再任意指定一张画，这个神经网络就会能将你的照片转换成那张指定的画的艺术风络。

 You can add the artistic style of Vincent van Gogh’s beautiful Starry Night to it and get some gorgeous results.

你可以将拍摄的那张照片转换成梵高的《星月夜》，成品很漂亮

Or, if you are looking for a bit more emotional or may I say disturbed look you can go for Edvard Munch’s The Scream ; for some stunning results.

要不然，如果你在寻找更情绪化、更扰乱的艺术风格的话，你可以选定蒙克的《尖叫》，成品也很漂亮。

And of course the mandatory Picasso. So as you can see deep neural networks are capable of amazing things and we expect even more revolutionary works in the very near future.

当然硬要转换成毕加索的艺术风格也是可行的。因此正如你所见，深度神经网络的潜力惊人，我们期待在不久的将来深度神经网络能带来更多革命性的应用

注：

论文地址：

https://arxiv.org/abs/1508.06576

复现论文代码地址：

https://github.com/jcjohnson/neural-style