I love how you love your job, you're a lucky guy

@OrangeC7

4 жыл бұрын

I love how you love how he loves his job

@ramseshendriks2445

4 жыл бұрын

@@OrangeC7 I love how you love how hes loves that Károly loves his job

@questforenlightenment441

4 жыл бұрын

Ramses Hendriks I love how you love that orangeC7 loves how alejandro loves how Karóly loves his job.

@gavinriley5232

4 жыл бұрын

@@questforenlightenment441 I love how you love that they love that they love that they love that he loves his job.

“And now squeeze that paper...” - the Hold On To Your Papers icon in the corner of the screen should change to Squeeze That Paper icon after those words.

@TwoMinutePapers

4 жыл бұрын

I was thinking about that too. We should do this!

@whatisthemaximumcharacterl7316

4 жыл бұрын

Wait but if we squeeze it, it gets crinkled.

I`m a vfx artist in Brazil, and I allways recommend this channel for everyone in my studio, thanks for the great content man ! keep it up

@TwoMinutePapers

4 жыл бұрын

You are very kind. Thank you so much! 🙏

These get better with each video and I get more excited with each video

@Danuxsy

4 жыл бұрын

you used "each video" twice in one sentence, stop doing that.

@deivisony

4 жыл бұрын

@@Danuxsy no u

I can do hair simulations easily, just with bald people

This AI instantly comments as soon as my videos are launched

@SeventhSolar

4 жыл бұрын

@天堂心臟 Games are an entirely different beast from physical systems. You can see OpenAI working on Starcraft II and Dota in videos on KZread. Their methods involve far less math tailored to a problem and far more focus on learning over time. Games are simply far too complex to describe in a mathematical model, and we have no idea what the ideal state would even look like for a game AI.

Amazing channel!

Me: Holding paper intensifies Also Me: Watch video Also Also Me: Squeezing paper intensifies

Amazing works. special one

So you're telling me a drone haircut really is possible?

Love that new logo!

Each new video from you shows an incredible advance in AI. I even think that maybe in like, two to four years, you will upload on a daily basis research on things that now seems like pure science fiction

@dtracers

4 жыл бұрын

What is most exciting is all the different ways these problems are being tackled.

It’s amazing that it can do all this automatically! I wonder what the output would be for really curly hair or braids.

I need this tech in my life.

Combine this with that ai that models someone's face

@omniscientomnipresent5500

4 жыл бұрын

I just thought that

@NaakaQ

4 жыл бұрын

before even that you could use something like that mobile app called 'FaceApp' to have a collection of similar looking versions of you that are old or young or whatever then make a 3d model of your face and use the other modified pictures to generate the different hair styles or coloration/whatever. your own clone army.

It's been three years and when can we finally use this in production?

This is the start of incredibly detailed VR avatar creation further on into the future

Man, this channel should be in schools. and have WAY more subscribers!

Imagine combining all these papers into one ai/physics engine!

@anonanon3066

4 жыл бұрын

Yeah. But then you would need an entire bakery to bake just a few frames of video physics

@kot3405

4 жыл бұрын

Actually not really

@dumbbass8867

4 жыл бұрын

You could also use this to randomly generate npcs on the spot, with thousands of voices to choose from and the ability to edit the pitch or volume could lead to many more. Combining that with the byped walking, and you've got a city with in depth characters in minuets!

@OrangeC7

4 жыл бұрын

So, basically what you're all saying, is to make Sword Art Online's World Seed?

And I'll see you nextimee..!!👍👌

Very cool article. Hope that one day, soon, we will se this in realtime game graphics

@VictorCampos87

4 жыл бұрын

That "our method" simulation takes 50ms to run 1 frame. So (1000 ms)/(50 ms/s) = 20 fps We are almost there.

Everytime you make a video on a subject I can't help but think of the application of these things in video games, naughty stuff or even propaganda. In the future, everything will be made possible, and more!

Hey Safay , can you predict in how many episodes you will present a paper that takes a complete Hollywood movie and extractor on full 3D for us to watch on any angle we want,,, 😜

This would work great for digital doubles in movies and TV. When you have to digitally replace an actor, like in a stunt for example.

should layer the simulated hair over the original image and see how well it blends.

how can I access the original papers discussed in the video?

*holding onto papers*

wow, I hope your youtube career goes well! Your videos are so good!.

Wow!

Could this be used to “preview” a haircut??

@iwanchandra3295

4 жыл бұрын

only need short video from your preferences

quick request - can you talk about the muzero results from deepmind..

Where do you find all these papers?!

What about curly hair or black power?

Is there any github code and weights available for this paper?

I like the progress, but I long for more, hair is usually a dead giveaway for AI. I am guessing there have been advancements with real time ray tracing helping hair refract light better?

This makes me uncomfortable as a 3D artist...

@powertomato

4 жыл бұрын

Fear not, the AI in 2D graphics is much more advanced than the one for 3D and yet artists are still around. As various AI features slowly make it into user software they just make artists more productive by automating some tedious work that had to be done manually.

@teaonabag

4 жыл бұрын

@@powertomato Yep, yep! Instead of being worried about your occupation, worry about having to look out for company mandated haircuts in your contract.

@yuhanzhang2882

4 жыл бұрын

it will just give you better tools to make animation

@WanderlustWithT

4 жыл бұрын

Adapt to change or get left behind , it's that simple.

Wow. With this, you could technically recreate a 3D human model from a single pictures. If there would be a working solution, I would love to use it.

wow!

@nearbygamerfanable

4 жыл бұрын

Nice

how much vram does this worth to render these actions

Can I somehow try those AIs by myself?

nice

Screw physics and optimization, we'll just have AI redraw each frame from what it learned from media.

What kind of application this can have other than in games?

@arnavrawat9864

4 жыл бұрын

It's a graphics application. Thus animated movies can be really realistic, easy and cheaper to make. Advertisements... Besides that rather than the graphics application, this trend of better graphics techniques will cause the betterment of ai reasearch field itself.

@precisiongaming8981

4 жыл бұрын

@@arnavrawat9864 Very well put, sir.

@jorge69696

4 жыл бұрын

In the future, hairdressers will have small computers with cameras that can take a picture or short video and instantly 3d model your entire head and face. From there, you can choose what haircut you want and can preview it in real time. Could also change hair color too. I predict it will be a phone app made by some Chinese or Russian company that steals all your facial data used for the rendering.

@xl000

4 жыл бұрын

@@jorge69696 yeahh .. hairdresser totally want to use computers to help them do their work. You know who was a hairdresser ? Heidi Gardner from SNL. Do you really see her tinkering with some software just to show a preview of what she may be able to do with your hair ? She .. or any hairdresser.. They're not nerds.

**CD projekt entered the chat**

So, how amazing is Aloy's hair going to be in Horizon Zero Dawn 2?

They should make something that captures body and hand poses from just one photograph! Surely that is easier than capturing hair geometry?!?

@0609Bhuwan

4 жыл бұрын

kzread.info/dash/bejne/iq2Mq6emkdO9m9o.html

@emf321

4 жыл бұрын

@@0609Bhuwan Amazing, thanks! That could lead to super simple motion capture techniques. Now we just need that as a pose-plugin for Daz Studio!

0:30 Is that a culturist wolf?

@wolfgangouille

4 жыл бұрын

Its clearly an armadillo.

will be video chat filters.

WHEN CAN WE HAVE ACESS TO THIS

Hi, I am not a scholar at all, just a young lad interested in those fields.. What can be the point of recreating the movement through video ? Would it be better to create the hair and then applying some physic engine instead of following it frame by frame ?

@Hexanitrobenzene

4 жыл бұрын

Physics simulations are always resource hungry, thus slow. Trained AI is capable of rendering very similar images with a lot less of resources. I don't know about the point, though. I assume it to be about creating a digital avatar from a short video or something. Seems like researchers in computer graphics are playing with the ideas at this time - raising problems and developing solutions, without serious applications, like a fundamental research in natural sciences... Or maybe I'm just out of imagination, as I'm far from computer graphics.

The paper only presents a few goal functions and a generic multi step neural network. No source code available. I'm always a bit skeptic when you can't reproduce their results without reimplementing the whole thing by yourself.

"Rotate us 75 degrees around the verticle, please." There is a shape change in this girl's hair, we can hypothesize Chris! It's actually happened - kzread.info/dash/bejne/ZXmrvLOdlMaZm7w.html

Jokes on you, I'm bald

We missed you

This could lead to video based invisibility.

1:45 creepy

Except afro type of hair

@kcfamilam5109

4 жыл бұрын

exactly my thought, even the "short curly hair" at 2:30 is pretty liberal with the classification. I would not call that curly hair at all lmfao.

@exosproudmamabear558

4 жыл бұрын

@@kcfamilam5109 Wavy was the term that should've been used, curly doesn't even come closer

@cachorro25

4 жыл бұрын

So what? This is about how strands of hair move, and for that you need long straight or wavy silky hair. Nappy hair don't apply

Wow so hopefully triple A games use this tech

Ezért fizetve vagyunk, de azért kell a szponzor pénze is, cuz reasons.

Killa

I am a bald guy - so it will be eaay

But why?

IMO this AI isn't perfect yet. It doesn't seem to be able to handle curled/permed hairs, model elasticity and friction between strands convincingly, and obviously its low sampling size that makes everyone looks like they're partially balding.

But it failed to capture any "curliness" at all?

@ronnetgrazer362

4 жыл бұрын

I felt equally underwhelmed. Maybe we're expecting too much, too soon.

Imagine implementing this is sexvilla or Honey collect

Karojonafachir

These footage are creepy af

robot r crepy

So we have deep fake, and with this man, we can now even replace you Karoly with read 3d Donald Trump model with perfect hair, his voice and gestures in the corner explaining other innovations in fluid simulations or neural networks.

I'm sure this deep learning technology could speak more humanly than you do. Anyway, great vid..

2 minute ads!

it looks terrible.