the challenge of dealing with data,

首先我先向大家介紹一個亟需解決的難題

data that we have to deal with

如何有效處理

in medical situations.

醫療過程中生成的

It's really a huge challenge for us.

海量數據.

And this is our beast of burden --

這些數據處理起來十分棘手.

this is a Computer Tomography machine,

這個正是解決此難題的關鍵.

a CT machine.

一台x光斷層掃描儀

It's a fantastic device.

即CT機.

It uses X-rays, X-ray beams,

這台機器非常先進.

that are rotating very fast around the human body.

它使用X線管發出X線束對患者進行掃描,

It takes about 30 seconds to go through the whole machine

同時這些X線管會圍繞患者高速旋轉.

and is generating enormous amounts of information

CT機完成一次掃描需要大約30秒

that comes out of the machine.

同時採集並輸出掃描所生成的

So this is a fantastic machine

大量信息.

that we can use

這台機器真的非常厲害

for improving health care,

它可以用來

but as I said, it's also a challenge for us.

提高衛生保健的質量.

And the challenge is really found in this picture here.

但正如之前所說,它也為我們帶來了一個問題.

It's the medical data explosion

從這張圖片大家可以看到這個問題.

that we're having right now.

它正是我們現在正在面臨的

We're facing this problem.

醫療數據爆炸.

And let me step back in time.

我們正在努力解決這個問題.

Let's go back a few years in time and see what happened back then.

讓我們先來回顧一下過去.

These machines that came out --

現在我們回到幾十年前

they started coming in the 1970s --

我要說的這些機器

they would scan human bodies,

於上世紀70年代開始投入使用

and they would generate about 100 images

醫生們用這些機器對患者進行人體掃描.

of the human body.

之後會生成大約100張

And I've taken the liberty, just for clarity,

人體影像.

to translate that to data slices.

恕我冒昧,為了方便理解,

That would correspond to about 50 megabytes of data,

我把這圖像轉換成等量的數據切片.

which is small

這些圖像大約相當於50MB的數據,

when you think about the data we can handle today

這個數量很小

just on normal mobile devices.

如果和現在我們每天打交道的信息量相比

If you translate that to phone books,

只與普通的移動設備相當.

it's about one meter of phone books in the pile.

如果拿電話簿的信息量做比的話,

Looking at what we're doing today

約相當於一米高的電話簿疊加.

with these machines that we have,

現在我們來看看今天

we can, just in a few seconds,

對這些機器的使用.

get 24,000 images out of a body,

現在,只需數秒

and that would correspond to about 20 gigabytes of data,

一位患者的人體掃描可以得到24,000張影像.

or 800 phone books,

這相當於 20 GB的數據,

and the pile would then be 200 meters of phone books.

800本電話簿.

What's about to happen --

壘起來大約有200米.

and we're seeing this; it's beginning --

然後呢,會發生甚麼?

a technology trend that's happening right now

我們可以看到,它已經開始了----

is that we're starting to look at time-resolved situations as well.

一種新的技術趨勢已經出現

So we're getting the dynamics out of the body as well.

我們開始考量時間分辨力.

And just assume

我們也需要得到書面化的診斷結果.

that we will be collecting data during five seconds,

現在我們假設

and that would correspond to one terabyte of data --

我們收集到了掃描5秒鐘所得數據,

that's 800,000 books

大約為1 TB.

and 16 kilometers of phone books.

相當於800,000本電話簿,

That's one patient, one data set.

壘起來約16千米.

And this is what we have to deal with.

這還只是掃描一個病患所得數據集.

So this is really the enormous challenge that we have.

這也正是我們需要處理的數據量

And already today -- this is 25,000 images.

所以說這個難題真的十分棘手.

Imagine the days

現在正是這個問題. 這裡有25,000張影像.

when we had radiologists doing this.

想像一下在以前

They would put up 25,000 images,

醫生們是這樣研究病理的.

they would go like this, "25,0000, okay, okay.

放到現在,他們要放25,000張圖像上去,

There is the problem."

到時候他們就要像這樣去看,”第25,000張,

They can't do that anymore. That's impossible.

哎,對,問題找到了.”

So we have to do something that's a little bit more intelligent than doing this.

現實不允許他們再這樣去找了.

So what we do is that we put all these slices together.

所以我們要想一些更聰明的辦法.

Imagine that you slice your body in all these directions,

我們得把這些切片再整合起來.

and then you try to put the slices back together again

我們先把自己沿這些各種方向切成數據片,

into a pile of data, into a block of data.

然後把這些切片再重新放回到一起,

So this is really what we're doing.

這樣一堆數據就成了一個數據塊

So this gigabyte or terabyte of data, we're putting it into this block.

這就是我們要做的.

But of course, the block of data

把這許多GB、TB的數據合成一個數據塊

just contains the amount of X-ray

當然,這個數據塊

that's been absorbed in each point in the human body.

只包含了在各個部分

So what we need to do is to figure out a way

被人體吸收了的X線束的數據.

of looking at the things we do want to look at

接下來我們要做的就是想個辦法

and make things transparent that we don't want to look at.

只顯示我們想看到那部分的數據,

So transforming the data set

隱去我們不想看到的部份.

into something that looks like this.

於是我們要把這個數據集

And this is a challenge.

變成這個樣子.

This is a huge challenge for us to do that.

這個不容易做到.

Using computers, even though they're getting faster and better all the time,

這個非常不容易做到.

it's a challenge to deal with gigabytes of data,

雖然現在電腦運行已越來越快且穩定

terabytes of data

利用電腦處理上GB 的數據,

and extracting the relevant information.

或者說上TB的數據

I want to look at the heart.

並提取出所需信息依然並不容易.

I want to look at the blood vessels. I want to look at the liver.

有時候要看一下心臟,

Maybe even find a tumor,

有時候要看一下血管,有時看肝臟,

in some cases.

也許有時候

So this is where this little dear comes into play.

要找一下看沒有腫瘤.

This is my daughter.

現在該我的小女兒出現了.

This is as of 9 a.m. this morning.

這就是我女兒.

She's playing a computer game.

這大概是今天上午9點.

She's only two years old,

她在玩電腦遊戲.

and she's having a blast.

她才只有兩歲,

So she's really the driving force

但是玩得很開心.

behind the development of graphics-processing units.

這樣的她正是

As long as kids are playing computer games,

催動圖像處理器進步的原動力.

graphics is getting better and better and better.

只要小孩子還在玩電腦遊戲,

So please go back home, tell your kids to play more games,

電腦圖像處理技術就會越來越好.

because that's what I need.

所以請大家回去告誡你們的小孩多玩遊戲吧,

So what's inside of this machine

我真的很需要這個.

is what enables me to do the things that I'm doing

我要說的是,我處理

with the medical data.

醫療數據要用的東西

So really what I'm doing is using these fantastic little devices.

就包含在這機器裡面.

And you know, going back

我要用到的就是這些能幹的小設備.

maybe 10 years in time

多年前

when I got the funding

大約十年前

to buy my first graphics computer --

我得到足夠的資金

it was a huge machine.

買了我的第一台繪圖電腦

It was cabinets of processors and storage and everything.

那台電腦體型非常大

I paid about one million dollars for that machine.

像塞滿處理器,存儲器等等等等的格子

That machine is, today, about as fast as my iPhone.

這台機器花了我大約一百萬美金.

So every month there are new graphics cards coming out,

現在這機器運行速度大概和我的iPhone一樣.

and here is a few of the latest ones from the vendors --

每個月都會有不同的新顯示卡面世.

NVIDIA, ATI, Intel is out there as well.

這是銷售商們推出的最新的顯示卡----

And you know, for a few hundred bucks

NVIDIA, ATI, 還有Intel.

you can get these things and put them into your computer,

只要花個幾百塊

and you can do fantastic things with these graphics cards.

就能買到這些裝到電腦裡面去,

So this is really what's enabling us

然後就可以做很多想做的事情.

to deal with the explosion of data in medicine,

要解決醫療數據爆炸的問題

together with some really nifty work

靠的正是這個.

in terms of algorithms --

再加上一些其他

compressing data,

邏輯運算之類的技術活----

extracting the relevant information that people are doing research on.

比如數據壓縮,

So I'm going to show you a few examples of what we can do.

以及提取醫生需要研究部分的信息.

This is a data set that was captured using a CT scanner.

接下來我為大家演示一下我們能做到的部分.

You can see that this is a full data [set].

這是使用CT機掃描時建成的一個數據集.

It's a woman. You can see the hair.

大家可以看到這是一套完整的數據.

You can see the individual structures of the woman.

這是一位女性. 從頭髮可以分辨出來.

You can see that there is [a] scattering of X-rays

大家可以看到這位女性身體各處的生理構造.

on the teeth, the metal in the teeth.

她牙齒上一塊散布的X線束,

That's where those artifacts are coming from.

那是牙齒上的一塊金屬.

But fully interactively

也就是人造物所在的地方.

on standard graphics cards on a normal computer,

然後只需

I can just put in a clip plane.

在裝有普通顯示卡的普通電腦上

And of course all the data is inside,

進行適當的編程,解析出一個剖面.

so I can start rotating, I can look at it from different angles,

當然所有的數據都沒在表面,

and I can see that this woman had a problem.

通過旋轉可以從不同的角度進行觀察,

She had a bleeding up in the brain,

我可以看到這位女性有一個問題,

and that's been fixed with a little stent,

她的大腦顱腔有一處出血,

a metal clamp that's tightening up the vessel.

醫生用一個支架和

And just by changing the functions,

一個金屬夾子夾緊血管來控制出血.

then I can decide what's going to be transparent

通過改變功能設置,

and what's going to be visible.

我可以決定讓哪部分隱去

I can look at the skull structure,

哪部分顯示出來.

and I can see that, okay, this is where they opened up the skull on this woman,

我可以只看他的頭骨部分,

and that's where they went in.

然後可以觀察出,哦,醫生是從這裡打開她的頭蓋骨,

So these are fantastic images.

然後是從這個地方著手進行手術.

They're really high resolution,

這些都是非常有用的圖像.

and they're really showing us what we can do

他們能提供非常有用的信息,

with standard graphics cards today.

能告訴我們今天用普通顯示卡

Now we have really made use of this,

我們能做些甚麼.

and we have tried to squeeze a lot of data

現在我們確實已經開始利用起這些顯卡,

into the system.

我們希望能利用這些顯卡處理

And one of the applications that we've been working on --

盡量多的數據.

and this has gotten a little bit of traction worldwide --

我們正在開發的一個應用----

is the application of virtual autopsies.

這個應用已經開始全球推廣----

So again, looking at very, very large data sets,

虛擬屍檢.

and you saw those full-body scans that we can do.

這一次,從這些信息量巨大的數據集中

We're just pushing the body through the whole CT scanner,

大家可以再一次看到全身掃描的使用.

and just in a few seconds we can get a full-body data set.

把屍體完全推進CT掃描儀,

So this is from a virtual autopsy.

只要數秒就可以得到一個全身數據集.

And you can see how I'm gradually peeling off.

這是一次虛擬屍檢的資料.

First you saw the body bag that the body came in,

大家可以看到一層一層的解構如何完成.

then I'm peeling off the skin -- you can see the muscles --

首先是覆著屍體的停屍袋,然後是屍體

and eventually you can see the bone structure of this woman.

然後剖開皮膚,出現肌肉

Now at this point, I would also like to emphasize

最後可以看到這位女性的骨骼結構.

that, with the greatest respect

在這裡,我要強調一下,

for the people that I'm now going to show --

對接下來我要展示的屍檢範例

I'm going to show you a few cases of virtual autopsies --

我是懷著極高的敬意.

so it's with great respect for the people

這些都是虛擬屍檢的應用案例

that have died under violent circumstances

我是懷著對死者最高的敬意

that I'm showing these pictures to you.

向大家展示

In the forensic case --

這些暴力死亡的屍檢案例.

and this is something

法醫屍檢裡

that ... there's been approximately 400 cases so far

近四年來

just in the part of Sweden that I come from

單就瑞典我所在的地區來說

that has been undergoing virtual autopsies

目前已經有大約400例

in the past four years.

法醫屍檢

So this will be the typical workflow situation.

採用了虛擬驗屍.

The police will decide --

虛擬驗屍的流程大概是這樣的.

in the evening, when there's a case coming in --

首先,大概在晚上

they will decide, okay, is this a case where we need to do an autopsy?

警方到達案發現場

So in the morning, in between six and seven in the morning,

根據現場情況決定是否需要虛擬驗屍.

the body is then transported inside of the body bag

之後大概在早上6點到7點,

to our center

警察們把屍體裝進停屍袋

and is being scanned through one of the CT scanners.

送到我們研究中心

And then the radiologist, together with the pathologist

使用CT機進行掃描.

and sometimes the forensic scientist,

接著放射性專家以及病理學專家

looks at the data that's coming out,

有時候再加上法醫學家

and they have a joint session.

共同研究

And then they decide what to do in the real physical autopsy after that.

從CT機得到的數據.

Now looking at a few cases,

由他們確定接下來實際屍檢的步驟.

here's one of the first cases that we had.

現在我們來看一些真實案例.

You can really see the details of the data set.

這是早期案例中的一個.

It's very high-resolution,

我們可以看到非常詳盡的數據,

and it's our algorithms that allow us

它們能提供很大的幫助.

to zoom in on all the details.

然後我們利用電腦的邏輯演算

And again, it's fully interactive,

可以對所有想看到的細節進一步放大.

so you can rotate and you can look at things in real time

這一次,同樣非常智能,

on these systems here.

這個系統中我們可以像實際屍檢一樣

Without saying too much about this case,

根據需要對屍體進行旋轉.

this is a traffic accident,

對這個案例無需做過多的描述.

a drunk driver hit a woman.

這是一起交通意外,

And it's very, very easy to see the damages on the bone structure.

司機醉酒駕駛,一名女性被撞.

And the cause of death is the broken neck.

大家可以很清楚看到骨架上所受創傷.

And this women also ended up under the car,

死因是頸骨骨折.

so she's quite badly beaten up

被害者當場死亡.

by this injury.

撞擊發生時,

Here's another case, a knifing.

死者受到重創.

And this is also again showing us what we can do.

這裡是另一個案子.一起持刀行凶案.

It's very easy to look at metal artifacts

這一次我們來看看可以發現甚麼.

that we can show inside of the body.

屍體體內的金屬人造物部分

You can also see some of the artifacts from the teeth --

非常明顯.

that's actually the filling of the teeth --

你還可以看到牙齒裡也有一些人造物,

but because I've set the functions to show me metal

那是補牙的填充物.

and make everything else transparent.

這裡我設定了只顯示金屬,

Here's another violent case. This really didn't kill the person.

其他被自動屏蔽.

The person was killed by stabs in the heart,

這是另一起暴力案件.這裡並不是致命傷.

but they just deposited the knife

真正死因是心臟被刺.

by putting it through one of the eyeballs.

後來兇手又把刀

Here's another case.

插進了被害人的眼睛.

It's very interesting for us

再來看另外一個案子.

to be able to look at things like knife stabbings.

能直觀看到諸如東西被刀刺破的樣子

Here you can see that knife went through the heart.

是很有意思的一件事情.

It's very easy to see how air has been leaking

這裡你可以看到心臟被刀刺穿.

from one part to another part,

可以很清楚看到空氣

which is difficult to do in a normal, standard, physical autopsy.

從一個部位漏往另一個部位.

So it really, really helps

這在常規屍檢中是很難觀察到的.

the criminal investigation

所以說,犯罪研究中,

to establish the cause of death,

虛擬屍檢可以幫助

and in some cases also directing the investigation in the right direction

判斷死者真實死因,

to find out who the killer really was.

以及必要時候幫助建立正確的

Here's another case that I think is interesting.

緝凶方向.

Here you can see a bullet

接下來也是一個很有意思的案子.

that has lodged just next to the spine on this person.

這裡可以看到有一顆子彈.

And what we've done is that we've turned the bullet into a light source,