TED2018
Supasorn Suwajanakorn: Fake videos of real people -- and how to spot them
蘇帕索恩蘇瓦耶納柯恩: 真人的假影片以及如何辨視出它們
Filmed:
Readability: 4.6
1,453,308 views
你認為你很擅長辨識出名人說出他們沒有說的話的那些假影片嗎？來看看這場驚人的演說和技術展示，了解怎麼辦到的。電腦科學家和研究生蘇帕索恩蘇瓦耶納柯恩說明他如何用人工智慧和 3D 建模，使影片與聲音同步，來創造出如相片般真實的假影片。他也進一步談到這項技術的道德意涵和可能創意，以及要採取什麼步驟，來避免它被誤用。
Supasorn Suwajanakorn - Computer scientist
Supasorn Suwajanakorn works on ways to reconstruct, preserve and reanimate anyone -- just from their existing photos and videos. Full bio
Supasorn Suwajanakorn works on ways to reconstruct, preserve and reanimate anyone -- just from their existing photos and videos. Full bio
Double-click the English transcript below to play the video.
00:12
Look at these images.
0
876
1151
看看這些影像。
00:14
Now, tell me which Obama here is real.
1
2051
2635
告訴我，當中哪一個
歐巴馬是真的？
歐巴馬是真的？
00:16
(Video) Barack Obama: To help families
refinance their homes,
refinance their homes,
2
4710
2861
（影片）歐巴馬：
協助家庭做房款再融資，
協助家庭做房款再融資，
00:19
to invest in things
like high-tech manufacturing,
like high-tech manufacturing,
3
7595
2647
投資於高科技製造業之類、
00:22
clean energy
4
10266
1159
乾淨能源
00:23
and the infrastructure
that creates good new jobs.
that creates good new jobs.
5
11449
2779
及能夠創造新工作機會的基礎建設。
00:26
Supasorn Suwajanakorn: Anyone?
6
14647
1484
講者：有人知道嗎？
00:28
The answer is none of them.
7
16155
1874
答案是：通通不是。
00:30
(Laughter)
8
18053
1114
（笑聲）
00:31
None of these is actually real.
9
19191
1786
這些通通不是真的。
00:33
So let me tell you how we got here.
10
21001
1840
讓我告訴各位，
我們如何走到這一步。
我們如何走到這一步。
00:35
My inspiration for this work
11
23940
1578
我做出這項作品的靈感
00:37
was a project meant to preserve our last
chance for learning about the Holocaust
chance for learning about the Holocaust
12
25542
5411
是一項計畫，計畫原來的目的是要
保存我們從大屠殺生還者那裡
保存我們從大屠殺生還者那裡
了解大屠殺的最後機會。
00:42
from the survivors.
13
30977
1768
00:44
It's called New Dimensions in Testimony,
14
32769
2627
它叫做「證詞的新維度」，
00:47
and it allows you to have
interactive conversations
interactive conversations
15
35420
3126
它讓你可以和大屠殺生還者的
00:50
with a hologram
of a real Holocaust survivor.
of a real Holocaust survivor.
16
38570
2556
立體投影進行互動式交談。
00:53
(Video) Man: How did you
survive the Holocaust?
survive the Holocaust?
17
41793
1966
（影片）男子：你如何
從大屠殺活下來？
從大屠殺活下來？
00:55
(Video) Hologram: How did I survive?
18
43783
1668
（影片）投影：我如何活下來？
00:57
I survived,
19
45912
1807
我能活下來，
01:00
I believe,
20
48419
1527
我相信，
01:01
because providence watched over me.
21
49970
3023
是因為有老天保佑我。
01:05
SS: Turns out these answers
were prerecorded in a studio.
were prerecorded in a studio.
22
53573
3454
講者：其實這些答案都是
在攝影棚預先錄影好的。
在攝影棚預先錄影好的。
01:09
Yet the effect is astounding.
23
57051
2452
但效果仍然很驚人。
01:11
You feel so connected to his story
and to him as a person.
and to him as a person.
24
59527
3619
你會覺得和他的故事
及他本人很有連結。
及他本人很有連結。
01:16
I think there's something special
about human interaction
about human interaction
25
64011
3301
我認為，人類互動中有著某種特性，
01:19
that makes it much more profound
26
67336
2757
它會讓使用者的體驗相當深切，
01:22
and personal
27
70117
2198
相當個人化，
01:24
than what books or lectures
or movies could ever teach us.
or movies could ever teach us.
28
72339
3485
而這些是書本、課程，或電影
沒有辦法教導我們的。
沒有辦法教導我們的。
01:28
So I saw this and began to wonder,
29
76267
2425
我看到這個之後，開始納悶，
01:30
can we create a model
like this for anyone?
like this for anyone?
30
78716
2810
我們能不能為任何人
創造這樣的模型？
創造這樣的模型？
01:33
A model that looks, talks
and acts just like them?
and acts just like them?
31
81550
2975
和他們有相似外表、
說話方式，和行為的模型？
說話方式，和行為的模型？
01:37
So I set out to see if this could be done
32
85573
2007
所以，我打算試試能不能辦到，
01:39
and eventually came up with a new solution
33
87604
2310
最終，我找出了一個新的解決方案，
01:41
that can build a model of a person
using nothing but these:
using nothing but these:
34
89938
3220
只要用下列這些東西，
就能建造出一個人的模型：
就能建造出一個人的模型：
01:45
existing photos and videos of a person.
35
93747
2214
一個人既有的照片和影片。
01:48
If you can leverage
this kind of passive information,
this kind of passive information,
36
96701
2617
若你能發揮這類被動資訊的功效，
01:51
just photos and video that are out there,
37
99342
2007
只要既有的照片和影片，
01:53
that's the key to scaling to anyone.
38
101373
2056
那就是能夠將規模擴大
到所有人的關鍵。
到所有人的關鍵。
01:56
By the way, here's Richard Feynman,
39
104119
1777
順道一提，這是理察費曼，
01:57
who in addition to being
a Nobel Prize winner in physics
a Nobel Prize winner in physics
40
105920
3413
他不只是諾貝爾物理獎的得主，
02:01
was also known as a legendary teacher.
41
109357
2453
也是一位傳奇性的老師。
02:05
Wouldn't it be great
if we could bring him back
if we could bring him back
42
113080
2198
如果我們能找他回來教課，
02:07
to give his lectures
and inspire millions of kids,
and inspire millions of kids,
43
115302
3265
鼓舞數百萬名孩子，不是很棒嗎？
02:10
perhaps not just in English
but in any language?
but in any language?
44
118591
2992
且不只是用英文，
還可以用任何語言？
還可以用任何語言？
02:14
Or if you could ask our grandparents
for advice and hear those comforting words
for advice and hear those comforting words
45
122441
4602
或是，我們的祖父母已經不在了，
還能問問他們意見，
還能問問他們意見，
聽他們說些撫慰的話，如何？
02:19
even if they're no longer with us?
46
127067
1770
02:21
Or maybe using this tool,
book authors, alive or not,
book authors, alive or not,
47
129683
3396
或許能用這樣的工具，
讓不論還活著或已故的作家
讓不論還活著或已故的作家
02:25
could read aloud all of their books
for anyone interested.
for anyone interested.
48
133103
2937
出來朗讀他們的書
給想要聽的讀者聽。
給想要聽的讀者聽。
02:29
The creative possibilities
here are endless,
here are endless,
49
137199
2437
這個工具有著無限的創意可能性，
02:31
and to me, that's very exciting.
50
139660
1713
我對此感到十分興奮。
02:34
And here's how it's working so far.
51
142595
2002
目前，它的運作方式如下：
02:36
First, we introduce a new technique
52
144621
1667
首先，我們先採用一項新技術，
02:38
that can reconstruct a high-detailed
3D face model from any image
3D face model from any image
53
146312
4572
它能從任何影像，重新建構出
一個人非常細節的 3D 面部模型，
一個人非常細節的 3D 面部模型，
02:42
without ever 3D-scanning the person.
54
150908
2119
不需要對他做實際的 3D 掃瞄。
02:45
And here's the same output model
from different views.
from different views.
55
153890
2642
這是從不同視角，輸出的類似模型。
02:49
This also works on videos,
56
157969
1502
它也能用在影片上，
02:51
by running the same algorithm
on each video frame
on each video frame
57
159495
2852
對每一格影片進行同樣的演算法，
02:54
and generating a moving 3D model.
58
162371
2222
產生出會動的 3D 模型。
02:57
And here's the same
output model from different angles.
output model from different angles.
59
165538
2772
這是從不同的角度，輸出的類似模型。
03:01
It turns out this problem
is very challenging,
is very challenging,
60
169933
2534
結果發現，這個問題非常有挑戰性，
03:04
but the key trick
is that we are going to analyze
is that we are going to analyze
61
172491
2525
但關鍵的技巧在於，我們得要
03:07
a large photo collection
of the person beforehand.
of the person beforehand.
62
175040
2966
事先分析同一個人的大量照片集。
03:10
For George W. Bush,
we can just search on Google,
we can just search on Google,
63
178650
2539
如果要做小布希總統，
我們用 Google 搜尋就可以了，
我們用 Google 搜尋就可以了，
03:14
and from that, we are able
to build an average model,
to build an average model,
64
182309
2499
這樣，我們就可以建立
一個平均的模型，
一個平均的模型，
03:16
an iterative, refined model
to recover the expression
to recover the expression
65
184832
3111
一個經過反覆運算、改良後的模型，
它能夠很精密地重現出表情，
它能夠很精密地重現出表情，
03:19
in fine details,
like creases and wrinkles.
like creases and wrinkles.
66
187967
2336
如皺褶和皺紋等。
03:23
What's fascinating about this
67
191326
1403
很炫的一點是，
03:24
is that the photo collection
can come from your typical photos.
can come from your typical photos.
68
192753
3423
只要用你一般的照片集
就可以做到了。
就可以做到了。
03:28
It doesn't really matter
what expression you're making
what expression you're making
69
196200
2603
你在照片中做什麼表情都無所謂，
03:30
or where you took those photos.
70
198827
1885
在任何地方拍的照片都行。
03:32
What matters is
that there are a lot of them.
that there are a lot of them.
71
200736
2400
重要的是，要有很多張照片。
03:35
And we are still missing color here,
72
203160
1736
這裡，我們還缺了顏色，
03:36
so next, we develop
a new blending technique
a new blending technique
73
204920
2348
所以，下一步，我們開發出
一種新的混和技術，
一種新的混和技術，
03:39
that improves upon
a single averaging method
a single averaging method
74
207292
2836
它是從單一平均化法改良而來的，
03:42
and produces sharp
facial textures and colors.
facial textures and colors.
75
210152
2818
能夠產生出精細的
面部質感和顏色。
面部質感和顏色。
03:45
And this can be done for any expression.
76
213779
2771
任何表情都可以使用。
03:49
Now we have a control
of a model of a person,
of a model of a person,
77
217485
2499
現在我們可以控制一個人的模型，
03:52
and the way it's controlled now
is by a sequence of static photos.
is by a sequence of static photos.
78
220008
3795
控制它的方式，是透過
一系列靜態照片來進行。
一系列靜態照片來進行。
03:55
Notice how the wrinkles come and go,
depending on the expression.
depending on the expression.
79
223827
3126
注意看，根據不同表情，
皺紋有時會出現有時會消失。
皺紋有時會出現有時會消失。
04:00
We can also use a video
to drive the model.
to drive the model.
80
228109
2746
我們也可以用影片來驅動模型。
04:02
(Video) Daniel Craig: Right, but somehow,
81
230879
2593
（影片）丹尼爾克雷格：
對，但不知何故
對，但不知何故
04:05
we've managed to attract
some more amazing people.
some more amazing people.
82
233496
3771
我們更能吸引到很了不起的人。
04:10
SS: And here's another fun demo.
83
238021
1642
講者：這是另一個有趣的展示。
04:11
So what you see here
are controllable models
are controllable models
84
239687
2246
各位在這裡看到的是可控制的模型，
04:13
of people I built
from their internet photos.
from their internet photos.
85
241957
2444
我用他們在網路上的照片建立的。
04:16
Now, if you transfer
the motion from the input video,
the motion from the input video,
86
244425
2904
如果把輸入影像的動作做轉換，
04:19
we can actually drive the entire party.
87
247353
2152
我們就可以讓所有人跟著動起來。
04:21
George W. Bush:
It's a difficult bill to pass,
It's a difficult bill to pass,
88
249529
2172
喬治布希：要通過
這個法案很困難，
這個法案很困難，
04:23
because there's a lot of moving parts,
89
251725
2303
因為有很多會變動的部分，
04:26
and the legislative processes can be ugly.
90
254052
5231
且立法過程有時是很醜陋的。
04:31
(Applause)
91
259307
1630
（掌聲）
04:32
SS: So coming back a little bit,
92
260961
1837
講者：先倒帶一下，
04:34
our ultimate goal, rather,
is to capture their mannerisms
is to capture their mannerisms
93
262822
3191
我們的終極目標，
其實是要捕捉他們的動作習性，
其實是要捕捉他們的動作習性，
04:38
or the unique way each
of these people talks and smiles.
of these people talks and smiles.
94
266037
3045
或是說每個人說話
和微笑的獨特方式。
和微笑的獨特方式。
04:41
So to do that, can we
actually teach the computer
actually teach the computer
95
269106
2313
為了做到這一點，我們要教導電腦
04:43
to imitate the way someone talks
96
271443
2222
去模仿一個人說話的方式，
04:45
by only showing it
video footage of the person?
video footage of the person?
97
273689
2420
做法是把那個人的影片給電腦看。
04:48
And what I did exactly was,
I let a computer watch
I let a computer watch
98
276898
2577
我所做的其實就是讓電腦觀看
04:51
14 hours of pure Barack Obama
giving addresses.
giving addresses.
99
279499
3277
十四小時的影片，
內容全是歐巴馬在演講。
內容全是歐巴馬在演講。
04:55
And here's what we can produce
given only his audio.
given only his audio.
100
283443
3516
只要提供他的聲音，
我們就能產生出這樣的成果。
我們就能產生出這樣的成果。
04:58
(Video) BO: The results are clear.
101
286983
1777
（影片）歐巴馬：結果很清楚。
05:00
America's businesses have created
14.5 million new jobs
14.5 million new jobs
102
288784
4349
在 75 個月內，
美國的企業已經創造出了
美國的企業已經創造出了
05:05
over 75 straight months.
103
293157
2774
1450 萬個新工作機會。
05:07
SS: So what's being synthesized here
is only the mouth region,
is only the mouth region,
104
295955
2905
講者：這裡合成的只有嘴部區域，
05:10
and here's how we do it.
105
298884
1540
我們的做法是這樣的。
05:12
Our pipeline uses a neural network
106
300764
1826
我們的指令傳遞途徑
使用的是類神經網路，
使用的是類神經網路，
05:14
to convert and input audio
into these mouth points.
into these mouth points.
107
302614
2936
將輸入的聲音轉換成那些嘴部的點。
05:18
(Video) BO: We get it through our job
or through Medicare or Medicaid.
or through Medicare or Medicaid.
108
306547
4225
（影片）歐巴馬：透過我們的運作
或醫療保險、醫療補助來達成。
或醫療保險、醫療補助來達成。
05:22
SS: Then we synthesize the texture,
enhance details and teeth,
enhance details and teeth,
109
310796
3420
講者：接著我們再合成肌理，
強化細節和牙齒，
強化細節和牙齒，
05:26
and blend it into the head
and background from a source video.
and background from a source video.
110
314240
3074
將它結合到頭部
以及來源影片的背景上。
以及來源影片的背景上。
05:29
(Video) BO: Women can get free checkups,
111
317338
1905
（影片）歐巴馬：
女性能接受免費檢查，
女性能接受免費檢查，
05:31
and you can't get charged more
just for being a woman.
just for being a woman.
112
319267
2968
你不會因為身為女性，
就要付比較多錢。
就要付比較多錢。
05:34
Young people can stay
on a parent's plan until they turn 26.
on a parent's plan until they turn 26.
113
322973
3306
年輕人在 26 歲前都可以
掛在一位家長的方案底下。
掛在一位家長的方案底下。
05:39
SS: I think these results
seem very realistic and intriguing,
seem very realistic and intriguing,
114
327267
2952
講者：我認為這些結果
非常真實且有意思，
非常真實且有意思，
05:42
but at the same time
frightening, even to me.
frightening, even to me.
115
330243
3173
但同時，連我也覺得它們很嚇人。
05:45
Our goal was to build an accurate model
of a person, not to misrepresent them.
of a person, not to misrepresent them.
116
333440
4015
我們的目標是要建造出很精確的
人類模型，而不是要故意歪曲他。
人類模型，而不是要故意歪曲他。
05:49
But one thing that concerns me
is its potential for misuse.
is its potential for misuse.
117
337956
3111
但我擔心的一件事情，
是這項工具可能被誤用。
是這項工具可能被誤用。
05:53
People have been thinking
about this problem for a long time,
about this problem for a long time,
118
341958
2971
從 Photoshop 剛出現
在市場上之後，
在市場上之後，
05:56
since the days when Photoshop
first hit the market.
first hit the market.
119
344953
2381
長久以來大家都一直
在想著這個問題。
在想著這個問題。
05:59
As a researcher, I'm also working
on countermeasure technology,
on countermeasure technology,
120
347862
3801
身為研究者，
我也在研究反制的技術，
我也在研究反制的技術，
06:03
and I'm part of an ongoing
effort at AI Foundation,
effort at AI Foundation,
121
351687
2942
我參與了人工智慧
基金會的一項計畫，
基金會的一項計畫，
06:06
which uses a combination
of machine learning and human moderators
of machine learning and human moderators
122
354653
3397
結合機器學習和人類調節
06:10
to detect fake images and videos,
123
358074
2144
來偵測出仿冒的影像和影片，
06:12
fighting against my own work.
124
360242
1514
及對抗我自己的作品。
06:14
And one of the tools we plan to release
is called Reality Defender,
is called Reality Defender,
125
362675
3190
我們打算要推出的工具之一，
叫做「真相守衛者」，
叫做「真相守衛者」，
06:17
which is a web-browser plug-in
that can flag potentially fake content
that can flag potentially fake content
126
365889
4039
它是瀏覽器的插件，
能標出有可能是仿冒的內容，
能標出有可能是仿冒的內容，
06:21
automatically, right in the browser.
127
369952
2533
全自動化，就裝在瀏覽器上。
06:24
(Applause)
128
372509
4228
（掌聲）
06:28
Despite all this, though,
129
376761
1453
不過，儘管有這些努力，
06:30
fake videos could do a lot of damage,
130
378238
1840
仿冒影片仍能造成許多傷害，
06:32
even before anyone has a chance to verify,
131
380102
3294
甚至在任何人有機會
驗證之前，傷害就已造成，
驗證之前，傷害就已造成，
06:35
so it's very important
that we make everyone aware
that we make everyone aware
132
383420
2722
所以，非常重要的是，
要讓大家都能意識到，
要讓大家都能意識到，
06:38
of what's currently possible
133
386166
2007
目前有什麼是可能的，
06:40
so we can have the right assumption
and be critical about what we see.
and be critical about what we see.
134
388197
3369
我們因此得以有正確的假設，
並對所看到的內容抱持批評的態度。
並對所看到的內容抱持批評的態度。
06:44
There's still a long way to go before
we can fully model individual people
we can fully model individual people
135
392423
5007
要能為個人完整地建模
還有很長的一段路要走，
還有很長的一段路要走，
06:49
and before we can ensure
the safety of this technology.
the safety of this technology.
136
397454
2786
也還要很多努力才能確保
這項科技的安全性。
這項科技的安全性。
06:53
But I'm excited and hopeful,
137
401097
1587
但我很興奮，也抱著希望，
06:54
because if we use it right and carefully,
138
402708
3539
因為如果我們能正確、
小心地使用它，
小心地使用它，
06:58
this tool can allow any individual's
positive impact on the world
positive impact on the world
139
406271
4309
這工具能讓任何人正面地影響世界，
07:02
to be massively scaled
140
410604
2190
大規模地影響，
07:04
and really help shape our future
the way we want it to be.
the way we want it to be.
141
412818
2742
且能真正能用我們希望的方式
來型塑我們的未來。
來型塑我們的未來。
07:07
Thank you.
142
415584
1151
謝謝。
07:08
(Applause)
143
416759
5090
（掌聲）
ABOUT THE SPEAKER
Supasorn Suwajanakorn - Computer scientistSupasorn Suwajanakorn works on ways to reconstruct, preserve and reanimate anyone -- just from their existing photos and videos.
Why you should listen
Can we create a digital avatar that looks, acts and talks just like our sweet grandma? This question has inspired Supasorn Suwajanakorn, a recent PhD graduate from the University of Washington, to spend years developing new tools to make it a reality. He has developed a set of algorithms that can build a moving 3D face model of anyone from just photos, which was awarded the Innovation of the Year in 2016. He then introduced the first system that can replicate a person's speech and produce a realistic CG-animation by only analyzing their existing video footage -- all without ever bringing in the person to a Hollywood capture studio.
Suwajanakorn is working in the field of machine learning and computer vision. His goal is to bring vision algorithms out of the lab and make them work in the wild.
Supasorn Suwajanakorn | Speaker | TED.com