This algorithm taught itself to animate a still photo

This algorithm taught itself to animate a still photo
17:34 29 Noyabr 2016
Ölkə mətbuatı
A- A+

A team of researchers at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) have created a deep-learning algorithm that is able to generate its own videos and predict the future of a video based on a single frame.
 
As detailed in a paper to be presented next week at the Conference on Neural Information Processing Systems in Barcelona, the CSAIL team trained their algorithm by having it watch 2 million videos which would last for over a year if played back to back. These videos consisted of banal moments in day to day life to better accustom the machine to normal human interactions. Importantly, these videos were found “in the wild,” meaning they were unlabelled and thus didn’t offer the algorithm any clues as to what was happening in the video.
 
Drawing from this video data set, the algorithm would attempt to generate videos from scratch that mimicked human motion based on what it had observed in the 2 million videos. It was then pitted against another deep-learning algorithm which tried to discriminate between the videos that were machine generated and those that were real, a method of training machines called adversarial learning.
 
“What we found in early prototypes of this model was the generator [network] would try to fool the other network by warping the background or having these unusual motions in the background,” Carl Vondrick, a PhD candidate at CSAIL and lead author of the paper, told Motherboard. “What we needed to give the model was the notion that the world is mostly static.”
 
To rectify this issue, Vondrick and his colleagues created a “two-stream architecture” which forces the generative network to render a static background while objects in the foreground moved. This two-stream model generated much more realistic videos, albeit short ones with really low resolutions. The videos produced by the algorithm were 64 x 64 and comprised of 32 frames (standard movies shoot at 24 frames per second which means these videos just over one second long), depicting things like beaches, trains stations, and the faces of new born babies (these are particularly terrifying).
 
While the ability to generate a second of video from scratch may not sound like much, this far surpasses previous work in the field which was only able to generate a few frames of video with much stricter parameters in terms of the content. The main pitfall of the machine generated videos is that the objects in motion in the video, particularly people, were often rendered as “blobs,” although the researchers still found it “promising that our model can generate plausible motion.”
 
Indeed, this motion was so plausible that when the researchers showed a machine generated video and a ‘real’ video to workers hired through Amazon’s Mechanical Turk and asked them which they found to be more realistic, they chose the machine generated videos about 20 percent of the time.
 
Beyond generating original videos, one of the more promising results of this work is the ability to apply it to videos and photos that already exist. When the researchers applied their deep-learning algorithm to a still frame, the algorithm was able to discriminate among objects in the photo and animate them for 32 frames producing “fairly reasonable motions” for the objects. To Vondrick’s knowledge, this is the first time that a machine has been able to generate multi-frame video from a static image.
 
This ability to anticipate the motion of an object or person is crucial to the future integration of machines in the real world, insofar as this will allow machines to not take actions that might hurt people or help people not hurt themselves. According to Vondrick it will also help the field of unsupervised machine learning, since this type of machine vision algorithm received all of its input data from unlabeled videos. If machines really want to get good at recognizing and classifying objects, they’re going to need to be able to do this without label data for every single object.
 
But for Vondrick, one of the most exciting possibilities contained in his research has little scientific or real-world value.
 
“I sort of fantasize about a machine creating a short movie or TV show,” Vondrick said. “We're generating just one second of video, but as we start scaling up maybe it can generate a few minutes of video where it actually tells a coherent story. We're not near being able to do that, but I think we're taking a first step.”


RƏYLƏR

BU KATEQORİYADAN DİGƏR XƏBƏRLƏR

Ölkə mətbuatında bugün »

Dini məsələləri bilmədiyi halda fitva vermək düzdür?

23 İyul 2017 20:52

"Panatinaikos" - “Qəbələ”: Hakimlər açıqlandı

23 İyul 2017 20:48

Ölüm nədir?

23 İyul 2017 20:47

Yapon dənizində Şimali Koreyaya məxsus sualtı qayıq gözlənilmədən fəallaşıb

23 İyul 2017 20:46

Tər qoxusu gəldiyi halda namaz qılmaq olarmı?

23 İyul 2017 20:42

Mərasimə ekstravaqant geyindi - FOTO

23 İyul 2017 20:32

Şıxov çimərliyində iyrənc olay – TURİSTİ PALTARSIZ QOYDULAR

23 İyul 2017 20:29

İranda qatar relsdən çıxdı: 7 yaralı

23 İyul 2017 20:28

Sumqayıtda güclü yanğın: 8 ailə çətin durumda - VİDEO

23 İyul 2017 20:22

Mingəçevir sakini evində 50-dən çox pişik saxlayır - VİDEO

23 İyul 2017 20:22

Hz. Əlinin qılıncı neçə başlı idi?

23 İyul 2017 20:22

Bərdədə bir nəfər Kür çayında batıb

23 İyul 2017 20:20

Avtobusda sürücü ilə sərnişinlər arasında mübahisə - səbəb kondisionerdir - VİDEO

23 İyul 2017 20:17

Bakıda turistin həyatını xilas edən polis danışdı - VİDEO

23 İyul 2017 20:17

Dizləriniz niyə xırçıldayır?

23 İyul 2017 20:17

Bu binada 400-ə yaxın ailə 3 ildir liftsiz qalıb - VİDEO

23 İyul 2017 20:17

Bakıda yeni dələduzlar peyda olub - ALDANMAYIN - VİDEO

23 İyul 2017 20:17

Dəridə qaşıntı nədən olur?

23 İyul 2017 20:12

Ağcabədidə pambıq sahələri susuz qalıb - VİDEO

23 İyul 2017 20:12

300-dən çox ailənin yaşadığı binaya isti su verilmir - VİDEO

23 İyul 2017 20:12

Qadın boksçularımız AVRO-2017-də 3 medal qazandı

23 İyul 2017 20:09

Tovuzda iki minik avtomobili toqquşub

23 İyul 2017 20:05

Xəzərdə batan 2 nəfər xilas edilib

23 İyul 2017 20:05

Alman gözəl Kannda dincəlir - FOTO

23 İyul 2017 20:05

Çoxallahlıq hansı dəlillərlə inkar olunur?

23 İyul 2017 19:57

Əfqanıstanda azı 34 "Taliban" üzvü məhv edilib

23 İyul 2017 19:55

Tovuzda iki avtomobil toqquşdu

23 İyul 2017 19:53

Kür çayında 1 nəfər batıb

23 İyul 2017 19:52

"Administrasiya Rusiyaya qarşı sanksiyaları dəstəkləyir" - Sara Sanders

23 İyul 2017 19:52

Fitva nə deməkdir?

23 İyul 2017 19:52

Hermofroditizm nədir?

23 İyul 2017 19:47

Məhəmməd peyğəmbər kimdir?

23 İyul 2017 19:42

Bakıda qəribə hadisə: avtomobilin şüşə tavanını çəngəl-bıçaqla deşdilər

23 İyul 2017 19:39

Əfqanıstanda azı 34 "Taliban" üzvü məhv edilib

23 İyul 2017 19:37

Rəfiqələr çimərlikdə əyləndilər - FOTO

23 İyul 2017 19:32

"Taliban"a AĞIR ZƏRBƏ: 34 ölü, 30 yaralı

23 İyul 2017 19:28

Parkda sevişən cütlük hücuma məruz qaldı -

23 İyul 2017 19:28

Peyğəmbər və səhabələri bütləri və pirləri necə dağıtdılar?

23 İyul 2017 19:22

Böyrək sindromu ilə hemorragik qızdırma

23 İyul 2017 19:17

Bərdədə 1 nəfər Kürdə batıb

23 İyul 2017 19:14

Allergiyanın əlamətləri nələr ola bilər?

23 İyul 2017 19:12

Rusiyanın hərbi gəmiləri Bakıya gəlib

23 İyul 2017 19:07

Eli Saabdan yeni kolleksiya - FOTO

23 İyul 2017 19:05

Dalğıclar dənizdə xəzinə tapdı -

23 İyul 2017 19:02

İndoneziyada güclü zəlzələ

23 İyul 2017 18:53

Ağdamın icra başçısı: "Ermənistan ordusu hər gün rayonun işğal altında olan ərazilərinə od vurur"

23 İyul 2017 18:52

Dünyanın ən sakitləşdirici musiqisi: Weightless

23 İyul 2017 18:52

Belə də olur: Nənə öz nəvəsini dünyaya gətirdi - FOTO

23 İyul 2017 18:47

Çində tikinti kranı aşıb -

23 İyul 2017 18:42

Bakıda marşrut avtobusu qəza törədib: xəsarət alan var

23 İyul 2017 18:39