I killed them all, not just the artists but the drawings and the paintings too - Nvidia AI turns any sketch into a photo-realistic image

[rcmaehl]

Sources:

ZDNet

TechCrunch

 

Summary:
Nvidia's new GauGAN modeling tool can sketches into photo-realistic images based on an existing network of images
 

Media:

 

Quotes/Excerpts:

Quote

Today at Nvidia GTC 2019, the company unveiled a stunning image creator. Users of the software are with just a few clicks able to sketch images that are nearly photo-realistic. The software will instantly turn a couple of lines into a gorgeous mountaintop sunset. Called GauGAN, the software is just a demonstration of what’s possible with Nvidia’s neural network platforms. It’s designed to compile an image how a human would paint, with the goal being to take a sketch and turn it into a photorealistic photo in seconds. GauGAN has three tools: a paint bucket, pen and pencil. Select the cloud object and draw a line with the pencil, and the software will produce a wisp of photorealistic clouds. But these are not image stamps. GauGAN produces results unique to the input. Draw a straight line and it will produce a bare trunk. Draw a bulb at the top and the software will fill it in with leaves producing a full tree. GauGAN is also multimodal. If two users create the same sketch with the same settings, random numbers built into the project ensure that software creates different results. To have real-time results, GauGAN has to run on a Tensor computing platform. Nvidia demonstrated this software on an RTX Titan GPU platform. Bryan Catanzaro, VP of Applied Deep Learning Research, stated that with some modifications, GauGAN can run on nearly any platform, including CPUs, though the results might take a few seconds to display. In the demo, the boundaries between objects are not perfect and the team behind the project states it will improve. Nvidia turned to 1 million images on Flickr to train the neural network. Most came from Flickr’s Creative Commons. In GauGAN, change the season and the leaves will disappear from the branches. Or if there’s a pond in front of a tree, the tree will be reflected in the water. Catanzaro hopes this software will be available on Nvidia’s new AI Playground. He sees tools like this being used in video games to create more immersive environments. Even in this limited demo, it’s clear that software built around these abilities would appeal to everyone from a video game designer to architects to casual gamers. The company does not have any plans to release it commercially, but could soon release a public trial to let anyone use the software.

 

My Thoughts:

Is this the end of the Art Degree as we know it? When everyone can be a masterful artist, is anyone truly one? I'll definitely be installing this once it's released however so I can test out drawing better than some 5th grade level scribble I normally do.

[TempestCatto]

dA is about to be filled with even more sham artists. Though I would like to know what it does to cartoon characters..

[minibois]

Huh, this is interesting, cool, but most of all; scary.

I mean, it has been pretty obvious for a long time that something like this would become a reality; but I  didn't exactly expect it in this form..

 

Also, I love the name for the program. That's a somewhat clever reference.

[paddy-stone]

Wonder what it would do with anime characters.. purely for science of course

[porina]

5 minutes ago, rcmaehl said:

Is this the end of the Art Degree as we know it? When everyone can be a masterful artist, is anyone truly one? I'll definitely be installing this once it's released however so I can test out drawing better than some 5th grade level scribble I normally do.

In a similar way that having all the best ingredients wont make you into a good chef, having the ability to create images in this way doesn't mean everyone will be an artist. This is just another tool in the box. It may enable more to unleash their potential, but there will always be room for out of the box thinking.

[TempestCatto]

7 minutes ago, ActuallyMediocreOC said:

Edit-re: @TempestCatto I can see it now, really bad MS Paint fursonas sloppily pasted onto a realistic looking back drop

That's already a thing tho..

 

Spoiler

That's one example I could quickly find. Shite like this is all over dA. Including pony OC's made in pony creator and pasted onto even weirder backgrounds.

[myselfolli]

This looks awesome and I'll def. be trying it once it's released publicly. It's kinda scary to see how far we've come though...

[Drak3]

3 minutes ago, TempestCatto said:

Including pony OC's made in pony creator and pasted onto even weirder backgrounds.

Dude  we don't need to know about your kinks.

 

/s

[Dissitesuxba11s]

11 minutes ago, rcmaehl said:

Is this the end of the Art Degree as we know it?

Serious question for the artists out there: Does an Art Degree affect your chance of employment? I always thought that your portfolio will matter more than any degree.

[TempestCatto]

3 minutes ago, ActuallyMediocreOC said:

@TempestCatto Yea you right... Maybe now they can use it to make stonesonas? Anyone that has an oc as a rock or tree or something might use it to be uber realistic as a character? 

Treesona's (this is actually a joke based off of something that happened in the show itself, but still relevant enough to be funny I guess ._.)

Spoiler

 

[Guest]

For some reason, I hate this.

[Mooshi]

This isn't an artist tool. It's a party trick for noobs.

 

Art has soul, this does not. People don't buy computer generated landscapes. For realism, that's what professional grade cameras are for. For art, machines can't replicate that. Bob Ross' happy little trees, man.

[DrDerp]

Ready for Uncanny Valley faces all over the internet?

Even Speedwagon is afraid!

[BuckGup]

It’s impressive but it’s not that far off from content aware auto fill in photoshop. The only difference is this uses a drawing from a human. Otherwise you can’t just draw a human and expect it to work. You would need to train the model for humans instead like they have done previously with the thisisntarealperson website

[Brooksie359]

2 hours ago, rcmaehl said:

Sources:

ZDNet

TechCrunch

 

Summary:
Nvidia's new GauGAN modeling tool can sketches into photo-realistic images based on an existing network of images
 

Media:

 

Quotes/Excerpts:

 

My Thoughts:

Is this the end of the Art Degree as we know it? When everyone can be a masterful artist, is anyone truly one? I'll definitely be installing this once it's released however so I can test out drawing better than some 5th grade level scribble I normally do.

Can this program make random blotches on a canvas for abstract art?

[poochyena]

2 hours ago, rcmaehl said:

Is this the end of the Art Degree as we know it? When everyone can be a masterful artist, is anyone truly one?

How is this getting rid of artists? This is just another tool for artists no different than a paint brush or electronic drawing tablet. It still requires skill and imagination to use.

[DezGalbie]

The scenes rendered look to me like the kind of thing you would have seen in a mid-90s detective show when they did their one episode about a software engineer working on an "advanced" virtual reality system who got murdered for his code.

Doesn't matter which show, they all did that episode at some point.

[seon123]

This makes me think nVidia lives 2 weeks ahead of us. 

[Syntaxvgm]

2 hours ago, TempestCatto said:

dA is about to be filled with even more sham artists. Though I would like to know what it does to cartoon characters..

With the rate the machine learning animation is progressing, deepfakes and voice cloning....I've been imagining a future where ever fanfiction writer wouldn't even have to be able to draw to make a comic, or spend that much time to animate. Imagine, every shitty edgy fanfic animated. What an amazing future. 

[rcmaehl]

7 minutes ago, Syntaxvgm said:

With the rate the machine learning animation is progressing, deepfakes and voice cloning....I've been imagining a future where ever fanfiction writer wouldn't even have to be able to draw to make a comic, or spend that much time to animate. Imagine, every shitty edgy fanfic animated. What an amazing future. 

I can assure you the worst already have fan movies.

[TempestCatto]

7 minutes ago, Syntaxvgm said:

With the rate the machine learning animation is progressing, deepfakes and voice cloning....I've been imagining a future where ever fanfiction writer wouldn't even have to be able to draw to make a comic, or spend that much time to animate. Imagine, every shitty edgy fanfic animated. What an amazing future. 

I don't have the skill to turn my fanfic's into comics, so I instead focused on making them more 'book like' and I think it makes them much more readable and just better in general. But I did have one other thought...think of the future of memes tho.

[Syntaxvgm]

56 minutes ago, rcmaehl said:

I can assure you the worst already have fan movies.

oh but Imagine anyone can do it with no effort any time with no ones help 

There will literally be 100s of these for every original series. 

[Bcat00]

Not gonna beat around the bush, I really want to try this out myself.

 

it looks so much fun

[TempestCatto]

I just took a closer look and I don't think those images it creates have a lot of depth to them. They look like a shitty photoshop job with lots of pixelation in most areas. Granted, it's creating something out of nothing so we can't really criticize too much.

[dtaflorida]

If it gets better it will find a use. I've seen wedding photography company have a room in back full of photoshoppers working to replace grey smoggy skies with clear blue and fluffy clouds from a stock photo library.