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

算法优化自拍 – 译学馆
未登陆,请登陆后再发表信息
最新评论 (0)
播放视频

算法优化自拍

Algorithmic Beautification of Selfies | Two Minute Papers #159

亲爱的学者粉丝们好呀 这里是 Károly Zsolnai-Fehér的两分钟论文
Dear Fellow Scholars, this is Two Minute Papers with Károly Zsolnai-Fehér.
今天我们要讨论一个严肃的科学话题
Today we’re going to talk about a rigorous scientific topic
那就是创作出完美的自拍图片
none other than the creation of the perfect selfie photo.
自拍当然是我们自己拍的 这就意味着这些是典型的短距照片
By definition selfies are made by us, which means that these are typically short-range photos,
由于相机镜头的透视变形 我们经常体验到
and due to the perspective distortion of the camera lens, we often experience unpleasant
一些令人不满意的效果 例如经过严重放大的鼻子和前额
effects like the heavy magnification the nose and the forehead.
那么 这项技术能够让我们拍摄照片 然后
And get this, this technique enables us to take a photo and after that, edit the perceived
在不改变任何其他的东西的情况下编辑相机的感应距离
camera distance for it without changing anything else.
通俗的说 算法美化!
Basically, algorithmic beautification!
这个算法的工作原理是这样的 先分析这个照片试图确定
This technique works the following way: we analyze the photo and try to figure out how
这个照片被拍摄时的镜头距离
distant the camera was when the photo was taken.
然后 我们创建一个这个相机视角的数字模型和一个面部的3D的模型
Then, we create a digital model of the perspective camera and create a 3D model of the face.
这个过程数学家喜欢称之为拟合
This is a process that mathematicians like to call fitting.
意思是 如果我们知道这个透视相机的光学系统我们可以从
It means that if we know the optics of perspective cameras, we can work backwards from the input
已有的照片输入开始处理 找出一个将会生成这个照片的合适设置
photo that we have, and find an appropriate setup that would result in this photo.
然后 我们可以调整镜头距离消除畸变
Then, we will be able to adjust this distance to even out the camera lens distortions.
但是还没完 因为我们有一个面部的数字模型 我们能做的还很多
But that’s not all, because as we have a digital 3D model of the face, we can do even more.
比如说 我们也可以将它转向多个方向
For instance, we can also rotate it around in multiple directions.
为了建立这样的一个3D模型 我们通常定位面部几个容易识别的热点
To build such a 3D model, we typically try to locate several well-recognizable hotspots
像 下巴 眉毛 鼻梁 鼻子下面的区域 眼睛
on the face, such as the chin, eyebrows, nose stem, the region under the nose, eyes and
嘴唇
lips.
然而 由于这些热点会得到一个很差的3D面部表现
However, as these hotspots lead to a poor 3D representation of the human face, the authors
作者又额外添加更多的热点到检测过程中
added a few more of these hotspots to the detection process.
这仍然只需不到五秒
This still takes less than 5 seconds.
之前 我们也讨论过关于一个基于神经网络的技术
Earlier, we also talked about a neural network-based technique that judged our selfie photos by
即通过给它一个分数来评判我们的自拍
assigning a score to them.
我非常希望看到那种技术将会如何评判通过本文这个技术
I would absolutely love to see how that work would react to a before and after photo that
美化前后的两张自拍
comes from this technique.
这样 我们将这个分数作为一个最大化问题
This way, we could formulate this score as a maximization problem, and as a result, we
最后我们会有一个自动化的技术 通过这些包含的操作真正的
could have an automated technique that truly creates the perfect selfie photo through these
创作完美自拍
warping operations.
最好的评价就是让现实成为我们的评判员
The best kind of evaluation is when we let reality be our judge, and use images that
用这些更近或更远的图片 去比较与这项技术的输出结果的不同
were taken closer or farther away and compare the output of this technique against them.
在本视频中这些真实的图像承载着背景真实的标签
These true images bear the ground truth label throughout this video.
这些不同经常是不可察觉的 而且为了更好的定位误差
The differences are often barely perceptible, and to provide a better localization of the
一些不同的图片被展示在这个论文中
error, some difference images are shown in the paper.
如果你对立体观察感兴趣 那里有一整章都是关于这的
If you are into stereoscopy, there is also an entire section about that as well.
作者同样也上传了一个他们这项成果的可交互的版本
The authors also uploaded an interactive version of their work online that anyone can try,
每个人都可以尝试 而且是免费的
free of charge.
那么像通常一样 非常欢迎你们在下面
So as always, your scholarly before and after selfie experiments are more than welcome in
评论你们的前后自拍实验
the comments section.
无论你是否已经订阅这个系列或者刚刚订阅
Whether you are already subscribed to the series or just subscribing now, which you
你都应该这么做 确保点击这个铃声图标绝不错过任何一集
should absolutely do, make sure to click the bell icon to never miss an episode.
我们有很多令人惊奇的成果在接下来的视频中展示出来
We have lots of amazing works coming up in the next few videos.
希望下次还能见到你
Hope to see you there again.
感谢你们的观看和慷慨的支持 下次见
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

本视频介绍了一个利用相机远近和热点定位优化自拍的算法

听录译者

收集自网络

翻译译者

【B】Cod

审核员

豆子

视频来源

https://www.youtube.com/watch?v=UjuBLS15JqM

相关推荐