Adobe today announced several new generative AI features for their applications in the Creative Cloud. Among them are Project Stardust which can remove objects in the background with AI, the Firefly update which can generate vectors and also Project Primrose which is clothing technology which can change patterns in the blink of an eye.
Additionally, Adobe also announced several generative AI features for video. First is Project Fast Fill which uses generative AI to add objects in the video. Interestingly, the objects added earlier will also be affected by the lighting changes that occur in the video.
Second is Project Scene Change which can combine two videos into one. For example, if you have a video of yourself in the office environment and a video of the scenery in an interesting tourist spot, these two videos can become one as if you were on vacation.
Adobe also has Project Res Up, where a video or animated GIF can be upscaled using AI. Old videos and GIFs can increase the resolution up to four times compared to the original source. In the demo shown the elephant video at SD resolution is upscaled to around 1440P via Res Up. It enhances the detail in the skin and water that cannot be seen in the original video.
Adobe also showed some new image generative AI features such as Project Draw and Delight where children's sketches can be turned into professional drawings. Project Glyph Ease that can change the typeface on existing graphics.
Project Poseable which can generate images and then use AI to change the position of the body such as the front view, hands raised up and so on. There are several 3D objects available on Poseble. Users only need to select the model and props they want to use as the basis for the generated image. Also possible is to take an existing photo and Poseable can then replicate the style in the uploaded photo with just one click.
Project Dub Dub Dub is a generative AI technology that can translate audio files from one language to another. At the same time can also produce a transcription of the translated text. In the demo on stage the languages supported by Dub Dub Dub are English, Spanish, French, German, Japanese, Mandarin, Jewish and Arabic. One drawback that Dub Dub Dub has is that the mouth movements are not synchronized with the generated audio.
Project See Through is a feature that allows AI to remove the effect of light reflection on any image. For example, a picture of a pet taken from outside a window, this picture can be as if there is no window.
Each of these new features will be offered in both Adobe Creative Cloud applications and the web version. For now most of these projects are beta, still in development or even concept and there is still no official date when they can be used by all.
In my opinion this is one of the very interesting features for those who are not very creative in the use of Photoshop and others. What do you think?