There are many users who use AI platforms to generate all kinds of images based on a description, images that we can later edit through Photoshop to make different adjustments that Artificial Intelligence has not offered us, such as modifying the smile, making bigger or smaller eyes and nose thanks, too, to AI. But if we want manipulate the main object, person or animalAt the moment, Photoshop offers us this option.
Google, MIT, the University of Pennsylvania in the United States, among others, have published the results of a project called DragGAN, a project that allows us to easily and quickly modify any image where we only have to select the object or part of the body of an animal or person and move it towards the direction we want it to have.
Thanks to this project, changing the position of a horse’s legs, opening a lion’s mouth and turning its head, modifying the size of an object or anything else that comes to mind is a matter of seconds and without It is necessary to have knowledge of photo editing, as is necessary if we want to edit images with Photoshop.
The operation of this AI is as simple as set the center of the pattern we want to edit and move it to where we want to find the result we are looking for as we can see in the image above. This process is not carried out immediately, but the AI must analyze the image in order to obtain the result we are looking for.
The possibilities offered by this AI, as we can see in the following video, are practically unlimited. At the moment, this is a project that is not publicly available, so we don’t have any chance to test it. Regarding the implementation of the operation of this AI, we also do not know what the plans of Google, MIT or the University of Pennsylvania are, but the number of possibilities that Google implements it in one of its platforms, such as Google Photos, is very high.
Drag Your GAN: Interactive Point-based Manipulation on the Generative Image Manifold
paper page: https://t.co/Gjcm1smqfl https://t.co/XHQIiMdYOA
May 19, 2023 • 07:04
This does not mean that in the future this function will not be available in other tools such as Photoshop, who probably has been working on a similar function for Photoshop for some time. A few weeks ago, Adobe introduced Firefly, its text-based image generator platform that allows us elements of an image in a simple and fast way. Companies like Nudify are already leveraging this type of technology.
The future that this point-based image modification AI presents to us is very encouraging within this field and offers practically infinite possibilities limited only to the imagination to obtain results that, for different reasons, we have not been able to capture or modify images. so that they adapt to what we are really looking for.