Researchers at MIT-IBM Watson AI Lab have put their heads together to develop an interesting AI tool that uses a neural network to produce images that are quite identical to real-life photographs that have originally been uploaded by its user.
After which, you can remove unwanted objects from the pictures or replace certain areas with new ones. This interesting AI tool is called GANPaint Studio and the demo version is already available at ganpaint.io.
The team behind this tool concedes that their system can potentially be used by those with untoward motives; however, they believe that more importantly, it could help various studies understand how Generative Adversarial Networks (GANs) perform to improve our ability at detecting fake photos.
The creators of the tool state, “The novelty we added for GANPaint Studio is that a natural image (of this category) can now be ingested and modified with semantic brushes that produce or remove units such as trees, brick-texture, or domes.
The demo is currently in low resolution and not perfect, but it shows that something like this is possible.”