GauGAN - Artist tool uses AI to create photorealistic images from sketches

[Deleted member 2197]

New tool (GauGAN) for artists creates photorealistic images from segmentation maps which are labeled sketches that depict the layout of a scene.

Artists can use paintbrush and paint bucket tools to design their own landscapes with labels like river, rock and cloud. A style transfer algorithm allows creators to apply filters — changing a daytime scene to sunset, or a photorealistic image to a painting. Users can even upload their own filters to layer onto their masterpieces, or upload custom segmentation maps and landscape images as a foundation for their artwork.

Live interactive demo link below.
https://www.nvidia.com/en-us/research/ai-playground/

Tool:
http://nvidia-research-mingyuliu.com/gaugan

 

[iroboto]

Beast. I wish I could even begin to contemplate how to approach this. The creativity in this field is amazing

 

[orangpelupa]

In the future, with final product, combined with VR could be like what this girl did

 

[Silent_Buddha]

Beast. I wish I could even begin to contemplate how to approach this. The creativity in this field is amazing

Looks like it just comes down to color fill (and perhaps screen position?) determining what type of stock imagery to use to fill in the space and what properties to assign to that space. Also set scenery to use for transitions between color fill areas.

Still pretty cool to see though.

Regards,
SB

 

[iroboto]

Looks like it just comes down to color fill (and perhaps screen position?) determining what type of stock imagery to use to fill in the space and what properties to assign to that space. Also set scenery to use for transitions between color fill areas.

Still pretty cool to see though.

Regards,
SB

it's doing a bunch of stuff, but it's starting off with basic premises, color helps determine what it is, but it's really the shape that is a bigger differentiator it appears.

 

[Deleted member 2197]

Will be interesting to see how far they can take this. Apparently it has some smarts since it will adjusts the whole landscape lighting depending on time of day, color or materials used in particular areas. Great idea for making your own desktop backgrounds.

 

[Shifty Geezer]

it's doing a bunch of stuff, but it's starting off with basic premises, color helps determine what it is, but it's really the shape that is a bigger differentiator it appears.

You tell it what to draw. You select bush, gravel, snow, etc and blob it in.

Here's Sky, Dirt, River, Mountain, Bush



I broke it with Snow and Sand in the same image!

 

[Shifty Geezer]

Looks like it just comes down to color fill (and perhaps screen position?) determining what type of stock imagery to use to fill in the space and what properties to assign to that space. Also set scenery to use for transitions between color fill areas.

You can select abstract templates as a reference image and it changes the results to make them abstract. The red hand, black blob image gets interesting results. It seems to be building some form of representation and chopping up a reference image.

 

[iroboto]

You can select abstract templates as a reference image and it changes the results t make them abstract. The red hand, black blob image gets interesting results. It seems to be building some for of representation and chopping up a reference image.

Okay that makes a lot more sense

 

[Deleted member 2197]

GANIMAL
We’ve all passed a Chihuahua on the street that’s the size of a guinea pig with the attitude of a German Shepherd. With GANimal, you can bring your pet’s alter ego to life by projecting their expression and pose onto other animals.

Once you input an image into the GANimal app, the image translation network unleashes your pet’s true self by projecting their unique characteristics onto everything from a lynx to a Saint Bernard.
https://www.nvidia.com/en-us/research/ai-playground/

 

[Deleted member 2197]

November 22, 2021
AI Art draws based on interpretation of phrase or sentence.