ADM-201 dump PMP dumps pdf SSCP exam materials CBAP exam sample questions

人工智能通过人脸照片合成3D模型 – 译学馆
未登录,请登录后再发表信息
最新评论 (0)
播放视频

人工智能通过人脸照片合成3D模型

AI Creates 3D Models From Faces | Two Minute Papers

亲爱的学霸们大家好 这里是由Károly Zsolnai-Fehér带来的两分钟论文
Dear Fellow Scholars, this is Two Minute Papers with Károly Zsolnai-Fehér.
将一个能够代表我们的角色 放置在数字电影或电脑游戏中 这该是多么酷啊?
How cool would it be to be able to place a character representing us in a digital film or a computer game?
当然 实现3D影像 以及数字化面部无光颜色贴图
Of course, it would clearly be an extremely laborious task to digitize the 3D geometry
是一项艰巨的任务
and the albedo map of our face.
无光颜色贴图的实质就是纹理 一种有颜色的图案
This albedo map means a texture, a colored pattern
它描述的是我们的皮肤是如何反射和吸收光线的
that describes how our skin reflects and absorbs light.
显然捕获这样的表现法是一种劳民伤财的做法
Capturing such a representation is clearly a very lengthy and expensive process.
所以我们提出了一个疯狂的想法:这项科技仅用一张简单的照片
So get this completely crazy idea: this technique creates this full digital representation of
就能创建出任何一张全数字化显示的人脸
any face from no more than one simple photograph.
只需要他们的一张照片 我们甚至可以在全球数据总量中找到出现过的历史人物
We can even get historical figures in our digital universe, all we need is one photograph of them.
那么现在 这些惊人的结果 简直亮瞎我的眼睛
And now, feast your eyes on these incredible results.
拍摄照片后 这种技术会创造出这样两张无光颜色贴图
After taking a photograph, this technique creates two of these albedo maps:
一张是完整的低频地图 能记录整个脸部
one is a complete, low frequency map, which records the entirety of the face,
但只能包括一些粗略的细节
but only contains the rough details.
另一张无光颜色贴图能显示出更微小的细节 但却并不完整
The other albedo map contains finer details, but in return, is incomplete.
还记得我们之前在某个系列中讨论过的纹理合成的方法吗?
Do you remember the texture synthesis methods that we discussed earlier in the series?
输入具有重复结构中的一小部分图像
The input was a tiny patch of image with a repetitive structure, and after learning the
在学习了这种结构的统计特性后 就有可能连续地使用了
statistical properties of these structures, it was possible to continue them indefinitely.
关键是 我们还可以举一反三 将这个不完整的无光颜色贴图
The key insight is that we can also do something akin to that here as well – we take this incomplete
尝试合成出缺少的细节部分
albedo map, and try to synthesize the missing details.
这想法简直太牛了!
Pretty amazing idea indeed!
作者引用了一种经典的学习算法
The authors of the paper invoke a classical learning algorithm,
使用一个卷积神经网络
a convolutional neural network to accomplish that.
我们使用的神经网络越深 出现在输出上的高频细节就会越明显
The deeper the neural network we use, the more high-frequency details appear on the
换句话说 我们得到的图像会更清晰
outputs, or in other words, the crisper the image we get.
在本文中 你会看到他们使用了众包进行的用户调研
In the paper, you will find a detailed description of their crowdsourced user study that was
来验证技术的详细描述 还包括了UI界面和答疑信息
used to validate this technique, including the user interface and the questions being asked.
这里还将其与我们之前所接触到的
There are also some comparisons against PatchMatch, one of the landmark techniques for texture
不同于PatchMatch的纹理合成标记技术进行对比
synthesis that we have also talked about in an earlier episode.
看到两分钟论文知识库内容日渐扩大 知识点联系更加紧密
It’s pretty amazing to see this Two Minute Papers knowledge base grow and get more and
真的太棒了
more intertwined.
我希望你们能和我一样乐在其中!
I hope you’re enjoying the process as much as I do!
当然 为响应大家的请求 现在可以购买到两分钟论文的T恤了!
Also, due to popular request, the Two Minute Papers T-shirts are now available!
这次我们使用了不同服务打印这些T恤
This time, we are using a different service for printing these shirts, please give us
请将您的意见反馈发给我们
some feedback on how you liked it.
我的e-mail地址会出现在视频说明中
I’ve put my e-mail address in the video description.
如果你附上一张自己穿着 两分钟论文印制的T恤的美照
If you attach a photo of yourself wearing some cool Two Minute Papers merch, we’ll be
我们会很高兴的
even more delighted!
打开两分钟论文的网址 你就可以直接进入
Just open two minute papers.com and you’ll immediately have access to it.
这条链接仅在美国国内提供运送服务
This link will bring you to the service that ships to the US.
需要从美国出口的链接 视频说明中会有
The link for shipping outside the US is available in the video description.
感谢您的收看与支持 我们下期再见!
Thanks for watching and for your generous support, and I’ll see you next time!

发表评论

译制信息
视频概述

通过PatchMatch的纹理合成方法,将照片合成3D的。

听录译者

收集自网络

翻译译者

[B]无牙无耳

审核员

知易行难

视频来源

https://www.youtube.com/watch?v=u9UUWqVquXo

相关推荐