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

表面几何细节的迁移 – 译学馆
未登录,请登录后再发表信息
最新评论 (0)
播放视频

表面几何细节的迁移

Geometric Detail Transfer | 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.
在三维建模领域 我们经常要去做一些并不是那么平滑的表面
In the world of digital 3D modeling, it often occurs that we are looking for surfaces that are not perfectly smooth,
而是有些表面细节的曲面
but have some sort of surface detail.
比如木桌上皱纹 雕刻 颗粒感等等就是我们想要加到模型上的细节的很好的实例
Wrinkles, engravings, grain on a wooden table are excellent examples of details that we can add to our models,
计算机图形学研究者们喜欢把他们统称为“置换贴图(displacement maps)”
and computer graphics people like to collectively call these things displacement maps.
艺术家们经常遇到这种情况,就是他们喜欢某个物体的表面细节
Artists often encounter cases where they like the displacements on one object,
而这个物体本身看起来并没什么意思
but the object itself is not really interesting.
然而 很有可能存在一些别的几何形状配上这些表面细节看起来就会很棒!
However, it could be that there is a different piece of geometry these details would look great on.
在这篇文章中 这个问题可以说已经被解决了 因为当你
Consider this problem solved, because in this piece of work,
输入两个三维模型 一个有吸引人的表面几何细节
the input is two 3D models: one with interesting geometric details,
而另一个就是我们想要把这些表面细节迁移上去的模型
and the other is the model onto which we transfer these surface details.
而输出就会是一个融合了这两个输入模型的三维几何形状
The output will be our 3D geometric shape with two of these models fused together.
文中的结果看起来碉堡了!
The results look absolutely amazing.
我现在就想把这种方法用到好几个项目里
I would love to use this right away in several projects.
这个算法的第一个关键部分就是“度量学习”
The first key part is the usage of metric learning.
等等 这就是个专业名词了 它是啥意思呢?
Wait, technical term, so what does this mean exactly?
“度量学习”是机器学习领域的一个经典技术
Metric learning is a classical technique in the field of machine learning,
用来解决那些我们想去学习的不同参数状态的距离 而这些距离又没有明确数学定义的情况
where we’re trying to learn distances between things where distance is mathematically ill-defined.
让我们讲的更简单些再举个例子 比如 我们有一个
Let’s make it even simpler and go with an example: for instance, we have a database
人脸数据库 我们想检索和输入人脸数据相似的数据
of human faces, and we would like to search for faces that are similar to a given input.
要实现这个目的 我们手工定义了一些距离度量 比如
To do this, we specify a few distances by hand, for instance,
我们可以定义长络腮胡的人和长小胡子的人有比较近的距离度量
we could say that a person with a beard is a short distance from one with a moustache,
和没有胡子的人就有比较远的距离度量
and a larger distance from one with no facial hair.
如果我们把足够多的包含这种距离度量的数据输入给机器学习算法
If we hand many examples of these distances to a learning algorithm,
它就能找出那些有相似胡须的人
it will be able to find people with similar beards.
在本视频所讲的这项工作中 “度量学习”被用来学习物体间的关系
And in this work, this metric learning is used to learn the relationship between objects
包括拥有丰富表面细节的和没有丰富的表面细节的模型
with and without these rich surface details.
而我们所说的“迁移”处理就受益于这种技术
This helps in the transferring process.
当我们想在一个新模型上创作一个新的表面纹理的时候 有几个障碍需要跨越
As to creating the new displacements on the new model, there are several hurdles to overcome.
第一 我们没法直接把表面纹理抓取下来放到一个不同的模型上
One, we cannot just grab the displacements and shove them onto a different model,
因为这些模型有可能看起来就不一样 有不一样的表面曲率 有不一样的大小等等
because it can potentially look different, have different curvatures and sizes.
而解决方法就是抓取这些表面细节的统计特性
The solution to this would be capturing the statistical properties of the surface details
并利用这些信息为目标模型合成新的表面细节
and use this information to synthesize new ones on the target model.
需要注意的是 我们无法像在二维图像里一样做这种纹理合成
Note that we cannot just perform this texture synthesis in 2D like we do for images,
因为当我们把合成的结果投影到三维模型上时
because as we project the result to a 3D model,
会使置换表面纹理产生严重的失真
it introduces severe distortions to the displacement patterns.
有点像把一个橡胶毯盖到一个形状复杂的物体上
It is a bit like putting a rubber blanket onto a complicated object.
橡胶毯的不同区域会有不同的扭曲
Different regions of the blanket will be distorted differently.
一定要看看文章中的相关部分 作者提供了更多的结果
Make sure to have a look at the paper where the authors present quite a few more results
当然对这种技术的复杂性也有很细致的描述
and of course, the intricacies of this technique are also described in detail.
希望这个算法的开源实现能快点出现
I hope some public implementations of this method will appear soon,
如果能马上用到的话想想就很兴奋
I would be quite excited to use this right away,
而且我确信一定有很多艺术家想要创作
and I am sure there are many artists who would love to create
这些超级细致的模型 给将来的动画电影和电脑游戏用
these wonderfully detailed models for the animated films and computer games of the future.
与此同时 我们有一个完全重构过的软件
In the meantime, we have a completely overhauled software
和一整套的硬件流程来制作这些视频
and hardware pipeline to create these videos.
我们把这些快乐而刺激的故事写在了Patreon(一个众筹网站)上
We have written down our joyful and perilous story of it on Patreon
如果你对我们视频的幕后制作感兴趣的话
– if you’re interested in looking a bit behind the curtain as to how these episodes are made,
一定要去看一看 视频说明中就有噢~
make sure to have a look, it is available in the video description.
感谢您的观看和支持 下次再见!
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

这次up主介绍了一篇表面几何细节迁移的文章,可以把一个模型的表面细节迁移到另一个不同的模型上

听录译者

收集自网络

翻译译者

GraphiCon-origamidance

审核员

知易行难

视频来源

https://www.youtube.com/watch?v=wz9cUncBdxw

相关推荐