Supasorn Suwajanakorn: Fake videos of real people -- and how to spot them
سوباسورن سواجاناكورن: فيديوهات مزيّفة لأشخاص حقيقيين...وكيفية تمييزها
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.
على إعادة تمويل منازلهم،
refinance their homes,
like high-tech manufacturing,
مثل التصنيع عالي التقنية،
that creates good new jobs.
على فرصتنا الأخيرة للتعلم عن محرقة اليهود
chance for learning about the Holocaust
interactive conversations
لشخصٍ حقيقي ناجٍ من المذبحة.
of a real Holocaust survivor.
survive the Holocaust?
قد تمّ تسجيلها مسبقاً في الأستوديو.
were prerecorded in a studio.
and to him as a person.
about human interaction
في التفاعل البشري
أو المحاضرات أو الأفلام.
or movies could ever teach us.
like this for anyone?
and acts just like them?
ويتصرّف مثلهم تماماً؟
using nothing but these:
باستخدام ما يلي فحسب:
this kind of passive information,
من المعلومات المستترة،
على جائزة نوبل في الفيزياء
a Nobel Prize winner in physics
if we could bring him back
ويلهم الملايين من الأطفال،
and inspire millions of kids,
but in any language?
و إنّما بأية لغة؟
على المشورة وسماع تلك الكلمات المطمئِنة
for advice and hear those comforting words
يمكن لمؤلفي الكتب، سواء كانوا أحياء أم لا،
book authors, alive or not,
for anyone interested.
لمن يهمه الأمر.
here are endless,
لوجه ثلاثي الأبعاد من أيّ صورة
3D face model from any image
للشخص.
من زوايا مختلفة.
from different views.
on each video frame
على كل مقطع فيديو
output model from different angles.
من زوايا مختلفة.
is very challenging,
is that we are going to analyze
of the person beforehand.
we can just search on Google,
يمكننا ببساطة البحث على غوغل
to build an average model,
to recover the expression
like creases and wrinkles.
مثل الطويات الجلدية واالتجاعيد.
can come from your typical photos.
من ضمن صورك العادية.
what expression you're making
that there are a lot of them.
a new blending technique
a single averaging method
facial textures and colors.
of a model of a person,
is by a sequence of static photos.
هي من خلال سلسلة من الصور الثابتة.
اعتماداً على تعبير الوجه.
depending on the expression.
to drive the model.
الفيديو لتحريك النموذج.
some more amazing people.
are controllable models
from their internet photos.
the motion from the input video,
It's a difficult bill to pass,
إنه مشروع قانون صعب التنفيذ
is to capture their mannerisms
of these people talks and smiles.
وابتسم بها كلّ من هؤلاء الأشخاص.
actually teach the computer
تعليم الكمبيوتر
video footage of the person?
I let a computer watch
وهو يلقي خطابات.
giving addresses.
given only his audio.
14.5 مليون وظيفة جديدة
14.5 million new jobs
is only the mouth region,
هو منطقة الفم فقط،
into these mouth points.
or through Medicare or Medicaid.
أو نظام الرعاية الطبية للعجزة أو الفقراء.
enhance details and teeth,
وتعزيز التفاصيل والأسنان
and background from a source video.
بالاعتماد على فيديو.
على معاينات مجانية
just for being a woman.
حتى بلوغ عمر 26.
on a parent's plan until they turn 26.
تبدو واقعية جدًّا ومثيرة للاهتمام،
seem very realistic and intriguing,
frightening, even to me.
من شخص ما، وليس تشويهه.
of a person, not to misrepresent them.
is its potential for misuse.
about this problem for a long time,
في الأسواق.
first hit the market.
على التكنولوجيا المضادة،
on countermeasure technology,
effort at AI Foundation,
في مؤسسة الذكاء الاصطناعي،
والمشرفين البشر
of machine learning and human moderators
is called Reality Defender,
تسمى "المدافع عن الواقع"
that can flag potentially fake content
يمكنه الإبلاغ عن محتوى مزيف محتمل
إلى الضرر
that we make everyone aware
and be critical about what we see.
ونكون حاسمين حول ما نراه.
we can fully model individual people
قبل أن نتمكن من نمذجة الأفراد بشكل كامل
هذه التكنولوجيا.
the safety of this technology.
لأي شخص في العالم
positive impact on the world
بالطريقة التي نريدها.
the way we want it to be.
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