Paul Debevec animates a photo-real digital face

翻译人员: Yin`ai Sun 校对人员: Halei Liu

One of the biggest challenges in computer graphics

计算机图形制作面临的巨大挑战之一

has been being able to create a photo-real,

是如何制作一个真实的

digital human face.

人物面部表情。

And one of the reasons it is so difficult is that, unlike aliens and dinosaurs,

不像外星人和恐龙，它之所以如此困难的原因是

we look at human faces every day.

我们每天都会看到人的面部表情

They are very important to how we communicate with each other.

面部表情对我们非常重要，因为我们通过表情来彼此沟通

As a result, we're tuned in to the subtlest things

由此可见我们会注意那些非常微妙的东西

that could possibly be wrong with a computer rendering,

通过计算机直接渲染出来的很可能是不正确的

in order to believe whether these things are realistic.

更无法使人相信这些是真实的

And what I'm going to do in the next five minutes

在接下来的5分钟里面

is take you through a process

我将带领大家一起浏览一个流程

where we tried to create a reasonably photo-realistic computer-generated face,

来通过计算机创造一个写实的人物表情

using some computer graphics technology we've developed,

通过我们开发的一系列计算机图形技术

and also some collaborators at a company called Image Metrics.

同时和一些来自Image Metrics公司的伙伴

And we're going to try to do a photo-real face

我们将尝试制作一个写实的

of an actress named Emily O'Brien, who is right there.

名叫艾米丽(Emily O'Brien)的女演员的表情，她就在那儿。

And that's actually a completely computer-generated rendering of her face.

那实际上是一个完全由计算机渲染生成的她的面部头像。

By the end of the talk, we're going to see it move.

在讲座的结尾，我们将看到她动起来

The way that we did this is we tried to start with Emily herself,

我们尝试的方式是通过Emily她自己

who was gracious enough to come to our laboratory

她非常优雅高尚地进入我们的实验室

in Marina Del Rey, and sit for a session in Light Stage 5.

在玛丽安德尔湾，并坐到了我们的设备灯光舞台5当中

This is a face-scanning sphere, with 156 white LEDs all around

这是一个面部扫描球体，有156个白色的发光二极管分部在其中，

that allow us to photograph her

配合我们对她进行拍照

in a series of very controlled illumination conditions.

通过一系列严格控制的光照条件

And the lighting that we use these days looks something like this.

我们近期使用的光照方式看起来是这样的。

We shoot all of these photographs in about three seconds.

我们在三秒钟之内拍摄出全部这些照片。

And we basically capture enough information

我们主要地会采集足量的信息

with video projector patterns that drape over the contours of her face,

通过图像投影模拟出她的面部轮廓

and different principle directions of light from the light stage,

以及通过灯光舞台得到的不同角度光照条件下的数据

to figure out both the coarse-scale

来计算出粗略的骨骼模型

and the fine-scale detail of her face.

以及她的精细的面部细节模型数据

If we zoom in on this photograph right here,

如果我们看放大的这张照片

we can see it's a really nice photograph to have of her,

我们可以很清晰地看到她的面部细节

because she is lit from absolutely everywhere at the same time

因为她的面部被同时从每个角度完全地照亮了

to get a nice image of her facial texture.

来得到一份非常高质量的她的面部表情贴图

And in addition, we've actually used polarizers on all the lights --

并且，我们实际上在所有灯上面应用了偏振镜

just like polarized sunglasses can block

就如同偏振的太阳镜可以阻挡

the glare off of the road,

马路上的耀眼光斑

polarizers can block the shine off of the skin,

偏振镜可以组织皮肤上出现光芒

so we don't get all those specular reflections to take this map.

所以我们没有采集到这些贴图的高光反射部分

Now, if we turn the polarizers around just a little bit,

现在，如果我们稍微调整一下偏振镜

we can actually bring that specular reflection

我们就可以得到高光的反射

of the skin back in,

在皮肤上面

and you can see she looks kind of shiny and oily at this point.

并且，你会看到她的面部看起来有些闪耀和油光。

If you take the difference between these two images here,

如果你取一下两张图像的差

you can get an image lit from the entire sphere of light

你得到的合成图像就是360度的光照条件下

of just the shine off of Emily's skin.

艾米丽(Emily)皮肤的光照效果

I don't think any photograph like this had ever been taken

我认为没有任何图像会这样拍摄

before we had done this.

在我们这样做之前。

And this is very important light to capture,

采集光照信息是非常重要的，

because this is the light that reflects off the first surface of the skin.

因为光照会反射出皮肤表面的最上一层

It doesn't get underneath the translucent

它不会出现在下层半透明层，

layers of the skin and blur out.

它会模糊掉。

And, as a result, it's a very good cue

由此可见，它是一个很好的线索

to the detailed shape of the skin-pore structure

来获得皮肤细致的毛孔结构

and all of the fine wrinkles that all of us have,

以及我们都有的皮肤褶皱细节的全部，

the things that actually make us look like real humans.

实际上，这些都使得它看起来更接近人类的真实效果。

So, if we use information that comes off of this specular reflection,

如果，我们使用来自高光反射的信息

we can go from a traditional face scan

我们可以通过传统的面部扫描而

that might have the gross contours of the face and the basic shape,

得到面部的粗略的轮廓和基本的形状，

and augment it with information

同时输入大量的信息

that puts in all of that skin pore structure and fine wrinkles.

皮肤的毛孔结构和纹理特征

And, even more importantly,

更重要的是，

since this is a photometric process that only takes three seconds to capture,

因为这是基于照片建模的流程而我们只采集了3秒钟的照片数据

we can shoot Emily

我们可以用一下午的时间

in just part of an afternoon,

拍艾米丽(Emily),

in many different facial poses and facial expressions.

拍摄不同的面部姿态和表情变化。

So, here you can see her moving her eyes around, moving her mouth around.

这里，你可以看到她转动眼睛，嘴巴的开合。

And these we're actually going to use to create a photo-real digital character.

这些，实际上都会被我们用来创造数字虚拟的写实人物。

If you take a look at these scans that we have of Emily,

来看一下这些我们对艾米丽(Emily)进行的扫描

you can see that the human face does an enormous amount of amazing things

你会看到我们人类面部表情是千变万化的

as it goes into different facial expressions.

在不同的面部表情变换之中

You can see things. Not only the face shape changes,

你会看到，不仅仅是面部形态发生变化

but all sorts of different skin buckling and skin wrinkling occurs.

以及各种各样的皮肤的弧线和皮肤上的褶皱出现。

You can see that the skin pore structure changes enormously

你可以看到皮肤毛孔结构也有着巨大的变化

from stretched skin pores

从被拉伸的皮肤毛孔，

to the regular skin texture.

到普通的皮肤质地

You can see the furrows in the brow and how the microstructure changes there.

你可以看到前额眉骨那里细微的变化。

You can see muscles pulling down at flesh to bring her eyebrows down.

你可以看到肌肉的收缩在她皱眉的时候。

Her muscles bulging in her forehead when she winces like that.

她前额的肌肉会这样的膨胀，在她畏惧的时候。

In addition to this kind of high-resolution geometry,

对于这样高分辨率的几何体，

since it's all captured with cameras, we've got a great texture map to use for the face.

因为这些都通过照相机采集，我们有足够的她的面部表情贴图。

And by looking at how the different color channels of the illumination,

通过颜色和亮度不同的通道，

the red and the green and the blue,

红绿蓝色彩通道

diffuse the light differently,

受光照不同的漫反射

we can come up with a way of shading the skin on the computer.

我们可以通过计算机总结出为皮肤添加阴影的方式。

Then, instead of looking like a plaster mannequin,

那么，不同于看起来如同石膏像的假人，

it actually looks like it's made out of living human flesh.

她看起来如同人类一样鲜活。

And this is what we used

这就是我们使用的

to give to the company Image Metrics

为Image Metrics公司

to create a rigged, digital version of Emily.

制作的写实的数字虚拟人物Emily。

We're just seeing the coarse-scale geometry here.

这里，我们看看一下粗糙的几何模型

But they basically created a digital puppet of her,

但他们已经对她创建了一些基本的数据程式，

where you can pull on these various strings,

你可以拖拽这些不同的坐标，

and it actually moves her face in ways that are

它实际上运动的方式就会依照

completely consistent with the scans that we took.

我们扫描出的数据那样。

And, in addition to the coarse-scale geometry,

并且，对于这个基本模型，

they also used all of that detail

他们也使用了全部的细节

to create a set of what are called "displacement maps"

来创建一些列的置换贴图

that animate as well.

同样可以对其进行动画设置。

These are the displacement maps here.

这里是一些置换贴图

And you can see those different wrinkles actually show up as she animates.

同时你可以看到不同的褶皱出现在她的动画之中。

So the next process was then to animate her.

所以，接下来的步骤是对她进行动画设置。

We actually used one of her own performances to provide the source data.

我们实际上应用她自身的表演作为源数据。

So, by analyzing this video with computer vision techniques,

通过计算机视觉技术来分析这些视频，

they were able to drive the facial rig

他们可以驱动表情的变化

with the computer-generated performance.

通过计算机的预置效果生成。

So what you're going to see now, after this,

所以，我们将看到，通过这一切处理，

is a completely photo-real digital face.

是一个完全写实的数字话面部。

We can turn the volume up a little bit if that's available.

我们可以开大一点声音，如果可以的话。

Emily: Image Metrics is a markerless, performance-driven animation company.

Emily:Image Metrics是一家追求高效的动画公司

We specialize in high-quality facial animation

我们致力于制作高质量的面部表情动画

for video games and films.

为视频游戏和电影。

Image Metrics is a markerless, performance-driven animation company.

Image Metrics是一家追求高效动画的公司。

We specialize in high quality facial animation

我们致力于制作高质量的面部表情动画

for video games and films.

为视频游戏和电影。

Paul Debevec: So, if we break that down into layers, here's that diffuse component

保罗·德贝维奇：如果我们把它分解到各个层，这是漫反射部分

we saw in the first slide.

我们在第一张幻灯片里面看到的。

Here is the specular component animating.

这是高光部分的动画。

You can see all the wrinkles happening there.

你们可以看到各处皮肤褶皱的出现。

And there is the underlying wireframe mesh.

这是原始的线框结构层。

And that is Emily herself.

那是Emily她自己。

Now, where are we going with this here?

现在，我们接下来会做些什么？

We've gone a little bit beyond Light Stage 5. This is Light Stage 6,

我们已经有了高于灯光舞台5的装置设备。这是灯光舞台6.

and we're looking at taking this technology

我们期望将这一技术

and applying it to whole human bodies.

应用到人的全身。

This is Bruce Lawmen, one of our researchers in the group,

这是布鲁斯 劳曼，我们研发团队中的一位，

who graciously agreed to get captured running in the Light Stage.

非常乐意配合我们采集他在灯光舞台中跑步的数据。

And let's take a look at a computer-generated version

下面让我们看一下计算机生成的

of Bruce, running in a new environment.

布鲁斯的影像，在一个全新的环境中跑步。

And thank you very much.

非常感谢。

(Applause)

（掌声）