May 22, 2023, 10:39 – Updated May 24, 2023, 09:12.
Adobe has been using AI algorithms in Photoshop for years: it has always been possible to remove objects or people or to change an image, but the new AI tool makes exciting new actions.
Knowledge The project is not such an app, but academic research by MIT and the University of Pennsylvania in collaboration with Google. The study shows how this new artificial intelligence tool called DragGAN — the concept of GAN, generative adversarial network, is almost an old data point in this regard — can help users modify images in ways never seen before.
Supervitamin Warp. In Photoshop the Warp tool (Deform) allows something similar, but with DragGAN that deformation goes much further because it does not deform as such, but rather allows images to, as seen in videos. as shared by @_akhaliqyou can make the head of a yawning lion or someone naturally laughing. Researchers have also posted some videos on the official document, but the success of the project has piled up on the top page and many of them have crashed.
It makes the soul sing. This is the only tool that doesn’t actually distort, but manipulates the actual objects. The application of these transformations gives absolutely coherent results, and means that from any original image we can obtain the desired effect in terms of position or expression, for example.
AI generates the image for you, AI handles it for you. A tool for generative AI models can be a strong complement like DALL-E 2, Diffusion Steady or Midjourney, especially since these generate a more spectacular first image, this tool demonstrates the ability to achieve the accuracy of that result. we asked
Simple interface. The tool itself offers a series of controls in its interface that will also demonstrate how simple the process is: it is enough to set a few principles, the direction in which the “deformation” will occur, and from there the program will take care of it. generate those final images.
Removable In the image above you can check the modifications of the images, which range from the gait of the horse or the position and the orientation of the head of the dog and cat to the model – which we can also rotate in 3D. or a real-time modification of a car that we can transform into a pickup van. The modifications in the landscapes are also amazing, and they allow, for example, that the lighting of the scene adapts to the position of the sun.
In Xataka | Adobe introduces AI-based tools that can detect if a face has been ‘photoshopped’ or remove ‘uhms’ from audio
