Photoshop with AI: The app impresses with big edits with ease

A new photo editor called DragGANIt was described in a scientific paper presented on the preprints platform on Thursday 18th using the GAN (Generative Adversarial Networks) learning model. arXiv. The tool allows the user to “drag” any point in the image with precise control over where the pixels go.

This means easily manipulating poses, shapes, expressions and layouts from countless categories. New tool that works similarly to Photoshop’s Warp, which adjusts the perspective of images in a simplified way. uses artificial intelligence to reconstruct a new underlying object, in addition to moving pixels.

Although the work has not been peer-reviewed and the product is not yet ready for consumption, images posted on Twitter and YouTube show that the images have been manipulated in various ways. read upside down, totally real. It’s arguable that AI technology can redefine what we mean by “photography”.

How does DragGAN work?

Combining cutting-edge effects like Photoshop’s Neural Filters (still in beta) with DxO PhotoLab’s U-Point identification technology, DragGAN does both in a much more user-friendly way.

in this way just point to the part of the image you want to edit and change the pose, expression and details of the subject in the photo with one click for impressively realistic results. There is also the possibility of more granular controls that make it possible to apply a mask to a particular area: for example, the face so that the algorithm only affects it.

The researchers are now preparing to “extend point-based editing to generative 3D models” in the near future. Until that time, The job of “retouching” photos will undergo a paradigm shift and approach “conversion”.. Something that might (logically) be questioned by some professional photographers, which won’t prevent the tool from becoming popular around the world.