Blaise Aguera y Arcas: Jaw-dropping Photosynth demo

翻译人员: Geng Luo 校对人员: dahong zhang

What I'm going to show you first, as quickly as I can,

首先，我要用最快的速度为大家演示

is some foundational work, some new technology

一些新技术的基础研究成果。

that we brought to Microsoft as part of an acquisition

正好是一年前，微软收购了我们公司，

almost exactly a year ago.

而我们为微软带来了这项技术，它就是Seadragon。

This is Seadragon, and it's an environment

Seadragon是一个软件环境，你可以通过它以近景或远景的方式

in which you can either locally or remotely interact

with vast amounts of visual data.

浏览浩瀚的可视化数据。

We're looking at many, many gigabytes of digital photos here

我们这里看到的是许多许多GB（千兆字节）级别的数码照片，

and kind of seamlessly and continuously zooming in,

对它们可以进行持续并且平滑的放大，

panning through it, rearranging it in any way we want.

可以通过全景的方式浏览它们，还可以对它们进行重新排列。

And it doesn't matter how much information we're looking at,

不管所见到的数据有多少、

how big these collections are or how big the images are.

图像集有多大以及图像本身有多大，Seadragon都拥有这样的处理能力。

Most of them are ordinary digital camera photos,

以上展示的图片大部分都是由数码相机拍摄的照片，

but this one, for example, is a scan from the Library of Congress,

但这个例子则不同，它是一张来自国会图书馆的扫描图片，

and it's in the 300 megapixel range.

拥有3亿个像素。

It doesn't make any difference

然而，浏览它并没有什么区别，

because the only thing that ought to limit the performance of a system like this one

因为限制系统性能的唯一因素只是：

你所使用的屏幕的像素数。

is the number of pixels on your screen at any given moment.

It's also very flexible architecture.

Seadragon同时也是一个非常灵活的架构。

This is an entire book, so this is an example of non-image data.

举个例子，这是一本完整的书，它的数据是非图像的(文本)。

This is "Bleak House" by Dickens.

这是狄更斯所著的《荒凉山庄》，一列就是一章的内容。

Every column is a chapter.

To prove to you that it's really text, and not an image,

我给大家证明一下这真的是文本而非图片，

we can do something like so, to really show

我们可以这样操作，

that this is a real representation of the text; it's not a picture.

大家可以看出这真的是文本，而不是一幅图片。

Maybe this is an artificial way to read an e-book.

也许这会是一种阅读电子书的方式，

I wouldn't recommend it.

但是我可不推荐这么做。

This is a more realistic case, an issue of The Guardian.

接下来是一个更加实际的例子，这是一期《卫报》。

Every large image is the beginning of a section.

每一张大图片是一版开篇，

And this really gives you the joy and the good experience

而报纸或者杂志的纸质版本本身就包含了多种比例的图片，

of reading the real paper version of a magazine or a newspaper,

在阅读的时候，读者会得到更好的阅读体验，

which is an inherently multi-scale kind of medium.

从而享受阅读的乐趣。

We've done something

我们在这里做了小小的改动

with the corner of this particular issue of The Guardian.

在这一期《卫报》得角上。

We've made up a fake ad that's very high resolution --

我们虚构了一个高分辨率的广告图片——

much higher than in an ordinary ad --

这比你平常看到的普通广告的分辨率要高很多，

and we've embedded extra content.

在图片中嵌入了额外的内容。

If you want to see the features of this car, you can see it here.

如果你希望看到这辆车的特性，你可以看这里。

Or other models, or even technical specifications.

你还能看到其他的型号，甚至技术规格。

And this really gets at some of these ideas

这种方式在一定程度上

about really doing away with those limits on screen real estate.

避免了屏幕实际使用面积的限制。

We hope that this means no more pop-ups

我们希望这个技术能够减少不必要的弹出窗口

and other rubbish like that -- shouldn't be necessary.

以及类似的垃圾信息。

Of course, mapping is one of those obvious applications

显然，对于这项技术的应用，

for a technology like this.

数字地图也是显而易见的应用之一。

And this one I really won't spend any time on,

对此，我真的不想花费太多的时间进行介绍，

except to say that we have things to contribute to this field as well.

我只想告诉大家我们已经对这个领域做出了自己的贡献。

But those are all the roads in the U.S.

这些只是在NASA的地理空间图片基础上

superimposed on top of a NASA geospatial image.

进行叠加处理而得到的美国的道路地图。

So let's pull up, now, something else.

现在，我们先放下这些，看看其他的。

This is actually live on the Web now; you can go check it out.

实际上，这项技术已经放到网上了，大家可以自己去体验一下。

This is a project called Photosynth, which marries two different technologies.

这个项目叫Photosynth，

它实际上融合了两个不同的技术：

One of them is Seadragon

一个是Seadragon，

and the other is some very beautiful computer-vision research

而另一个则是源自华盛顿大学的研究生Noah Snavely

done by Noah Snavely, a graduate student at the University of Washington,

所进行的计算机视觉研究的成果。

co-advised by Steve Seitz at U.W.

这项研究还得到了华盛顿大学Steve Seitz

and Rick Szeliski at Microsoft Research.

和微软研究院Rick Szeliski的协助。这是一个非常漂亮的合作成果。

A very nice collaboration.

And so this is live on the Web. It's powered by Seadragon.

这个项目在互联网上已经得到应用了，它是基于Seadragon技术构建的。

You can see that when we do these sorts of views,

你可以看到，我们轻松地对图片进行多种方式的查看，

where we can dive through images

从而能够对图片进行细致的剖析

and have this kind of multi-resolution experience.

并且拥有多分辨率的浏览体验。

But the spatial arrangement of the images here is actually meaningful.

不过，这些图片在三维空间的排列事实上是非常有意义的。

The computer vision algorithms have registered these images together

计算机视觉算法将这些图片联系到一起，

so that they correspond to the real space in which these shots --

那么这些图片就能够将真实空间呈现出来了，

all taken near Grassi Lakes in the Canadian Rockies --

而我们正是在这个空间里拍下了上述的照片——这些照片都是在

all these shots were taken.

加拿大落基山脉的格拉西湖（Grassi Lakes）附近拍下的——（所有照片）都是在这里拍下的。

So you see elements here

of stabilized slide-show or panoramic imaging,

因此你可以看到这里的元素是稳定的幻灯放映或者全景成像，

and these things have all been related spatially.

而这些内容在空间上都是关联的。

I'm not sure if I have time to show you any other environments.

我不确定我们是否有时间来展示更多的环境全景。

Some are much more spatial.

有很多例子比这个的空间感还要强。

I would like to jump straight to one of Noah's original data-sets --

下面让我们来看一下去年夏天，

this is from an early prototype that we first got working this summer --

我们利用Noah早期的数据库之一

所Photosynth的初期模型的建立。

to show you what I think

我认为

is really the punch line behind the Photosynth technology,

这可谓是我们这项技术的最抢眼之处。

It's not necessarily so apparent

这项技术不单单像我们在

from looking at the environments we've put up on the website.

网站上展示得那么简单明了。

We had to worry about the lawyers and so on.

主要因为我们制作网站时，要顾及到很多法律问题。

This is a reconstruction of Notre Dame Cathedral

这里是利用Flickr网站上

that was done entirely computationally from images scraped from Flickr.

的图像重建的巴黎圣母院。

你所要做的只是在Flickr网站上输入“巴黎圣母院”

You just type Notre Dame into Flickr,

and you get some pictures of guys in T-shirts, and of the campus and so on.

然后便能看到很多图片，包括留影的游人等等。

所有这些橘黄颜色的锥形都代表了一张

And each of these orange cones represents an image

that was discovered to belong to this model.

用来建立模型的图片。

And so these are all Flickr images,

这些全部是来自Flickr的图片，

and they've all been related spatially in this way.

被这样在空间里被串联起来。

We can just navigate in this very simple way.

接着，我们便可如此自如的进行浏览。

(Applause)

（鼓掌）

(Applause ends)

You know, I never thought that I'd end up working at Microsoft.

说实话，我从来没想过我会最后来为微软工作

It's very gratifying to have this kind of reception here.

受到这样欢迎，真挺令人高兴的。

(Laughter)

（笑声）

I guess you can see this is lots of different types of cameras:

我想你们可以看出

这些图片原自很多不同的相机：

it's everything from cell-phone cameras to professional SLRs,

从手机摄像头到专业单反。

quite a large number of them, stitched together in this environment.

如此大量的不同质量的照片，全被在这个环境下

拼合在了一起

If I can find some of the sort of weird ones --

让我来找些比较诡异的图片。

So many of them are occluded by faces, and so on.

看，不少照片包含了游客的大头照等等。

Somewhere in here there is actually a series of photographs -- here we go.

我记得这儿应该有

一个系列的照片 - 啊，在这儿。

This is actually a poster of Notre Dame that registered correctly.

这个是巴黎圣母院的海报。

We can dive in from the poster

我们可以钻到海报里

to a physical view of this environment.

去看整个重建的环境。

What the point here really is

这里的重点呢便是我们可以

is that we can do things with the social environment.

有效地利用网络社区。我们可以从每个人那里得到数据

This is now taking data from everybody --

from the entire collective memory, visually, of what the Earth looks like --

将每个人对不同环境

的记忆收集在一起，

and link all of that together.

共建成模型。

Those photos become linked, and they make something emergent

当所有这些图片交织在一起时，

所衍生出的

that's greater than the sum of the parts.

要远远超过单单收集起全部。

You have a model that emerges of the entire Earth.

这个模型所衍生出的，是整个地球。

Think of this as the long tail to Stephen Lawler's Virtual Earth work.

这如同是Stephen Lawler的《虚拟地球》的长尾市场。（Stephen Lawler 微软Virtual Earth项目主管）（见Long tail 长尾市场 TED: Chris Anderson ）

And this is something that grows in complexity as people use it,

这类模型，会随着人们的

使用而不断变的复杂，

and whose benefits become greater to the users as they use it.

变得更加有价值。

Their own photos are getting tagged with meta-data that somebody else entered.

用户的照片，会被大家

注上标签。

If somebody bothered to tag all of these saints

如果有人愿意为所有这些圣母院里的圣贤注上标签，

and say who they all are, then my photo of Notre Dame Cathedral

表明他们是谁，那我们的圣母院照片便会

suddenly gets enriched with all of that data,

一下子丰富起来，

and I can use it as an entry point to dive into that space,

然后呢，我们便能以这张照片为起点，进入这个空间，

into that meta-verse, using everybody else's photos,

这个由很多人的照片所搭建的虚拟世界，

and do a kind of a cross-modal

从而得到一种跨越模型，

and cross-user social experience that way.

跨越用户的交互体验。

And of course, a by-product of all of that is immensely rich virtual models

当然了，这一切所带来另外一个宝贵产物便是

一个非常丰富的模型 - 充斥

of every interesting part of the Earth,

这地球每个角落里有趣的景观。这些景观不再

collected not just from overhead flights and from satellite images

局限于航空和卫星图片，

and so on, but from the collective memory.

而是实实在在的人们按下快门一刻所收藏的记忆的集合。

Thank you so much.

非常感谢！

(Applause)

（掌声）

(Applause ends)

Chris Anderson: Do I understand this right?

Chris Anderson: 如果我理解正确的话，你们的这个软件将能够

What your software is going to allow,

is that at some point, really within the next few years,

在未来的几年内

all the pictures that are shared by anyone across the world

将来自全球的图片

are going to link together?

接合在一起？

BAA: Yes. What this is really doing is discovering,

BAA:是的。这个软件的真正意义便是去探索。

creating hyperlinks, if you will, between images.

它在图片间构建起超链接。

It's doing that based on the content inside the images.

这个接合的过程

完全是基于图片的内容。

And that gets really exciting when you think about the richness

更令人兴奋的

of the semantic information a lot of images have.

在于图片所包含的大量文字语义信息。

Like when you do a web search for images,

比如，你在网上所以一张图片，

you type in phrases,

键入关键词后，网页上的文字内容

and the text on the web page is carrying a lot of information

将包含大量与这个图片相关的信息。

about what that picture is of.

What if that picture links to all of your pictures?

现在，假设这些图片全都与你的图片相连，那将会怎样？

那时，所以这些语义信息的相互链接

The amount of semantic interconnection and richness

以及内容量将是

that comes out of that is really huge.

巨大的。这将是非常典型的网络效应。

It's a classic network effect.

CA: Truly incredible. Congratulations.

CA:Blaise，太难以置信了。祝贺你们！

BAA:非常感谢各位！