Are you an artist or a designer trying to make
quick tweaks to visuals and identifying fake images?
Say hello to the GANpaint Studio system
developed by a team from MIT and IBM.
Take a look at the apps on your phone.
I bet you can easily apply filters and enhance the
quality of photos with just a few clicks.
What if the same set of tools, driven by AI,
can create scenes from scratch?
That’s what GANpaint Studio can do for you…
Automatically generate realistic photographic images,
and edit objects inside the images.
That’s not all…
It can also help identify fake images.
GAN stands for generative adversarial network.
Using this online tool, you can upload an image and modify
several attributes such as size and appearance.
How about a real example?
We took one of the photos from the tool itself and
added more volume to the tree (image on right)
The process took a few clicks!
If you are a designer or an artist,
you will be amazed by the speed of this tool.
Computer-graphics editors can use the Studio
to edit video clips in less time.
GANpaint can help researchers better understand
neural networks and their underlying structures.
The system is intelligent too.
It wouldn’t put, for example, a window in the sky.
In other words, it understands the relationships between objects.