Photoshop’s “Generative Expand” feature uses AI to easily and convincingly expand images, combining real and imagined elements

[One of the new features in Adobe’s Photoshop image editing software lets users easily expand the contents of an image beyond the original size and shape, making it that much easier to combine actual and artificial elements in a photo and that much harder for viewers to distinguish between them. The short story below from XDA provides the basic information about “Generative Expand” and Adobe provides details in its blog post and a 3:38 minute video available on YouTube. ZDNet has a detailed, example-filled description of the feature along with Content Aware Fill and Generative Fill; the story also notes an advantage of using Photoshop:

“Adobe only used images in its own Adobe Stock collection, all of which it owns and licenses as part of your Creative Cloud license. So you’re very safe if you use a Photoshop-generated image and very exposed if you use something from one of the other AIs.”

TechCrunch notes that “Adobe says that it’s implemented filters to prevent Generative Expand from generating toxic content — a notorious problem for generative art AI.” Engadget is among the minority of press outlets that acknowledges the potential dangers of the feature and those like it:

“As with Generative Fill, Expand is meant primarily for creators who are more interested in artistic expression than accuracy. The catch, of course, is that you can also distort or exaggerate images. You can create a dramatic panorama of a pristine beach when the reality is far uglier, for instance. While manipulating images in editors like Photoshop is nothing new, the AI makes it easier to spread misinformation or otherwise mislead viewers.”

See also “AI-Generated Attempt To Reveal ‘The Rest Of The Mona Lisa’ Goes Viral” from Forbes. –Matthew]

[Image: Before (L) and after (R) using Generative Expand. Credit: Terry White]

Photoshop’s generative AI feature can now expand images seamlessly

Photoshop users can now utilize generative AI to expand or reduce images with or without text prompts.

By kKishalaya Kundu
July 27, 2023

Adobe integrated its Firefly family of generative AI models into the beta version of Photoshop in May, enabling users to add, expand, or remove content from images non-destructively, using simple text prompts. The company is now building on those upgrades by introducing a new feature called ‘Generative Expand’ that lets users expand and resize images seamlessly.

To resize images using generative AI, Photoshop users need to first click and drag the Crop tool to expand the canvas. Then just clicking on the ‘Generate’ button in Photoshop’s contextual taskbar will fill the new blank space with AI-generated content that blends seamlessly with the existing image. According to Adobe, the feature would come in especially handy if the subject is partially cut off, the image isn’t the required aspect ratio, or an object in focus is misaligned with other parts of the image.

The best part about Generative Expand in Photoshop is that it can be used to add content to any image with or without a text prompt. In case there’s no prompt, Photoshop will simply fill in the expanded canvas with AI-generated content that blends with the existing image. When using a prompt, the expanded content will always include the element(s) mentioned by the user. “With Generative Expand, you can spend less time editing and more time experimenting and adapting your images for your own creative needs”, said Adobe about its latest AI feature.

In addition to Generative Expand, Adobe also announced another development that is sure to help many creators around the world. Starting today, Firefly-powered features in Photoshop will now support text prompts in more than 100 languages. This is a welcome change for many creators around the world, as until now, the generative AI text prompts only worked with languages like English, etc.

In case you’re interested in trying out Generative Expand, you’ll need to update your Photoshop app, as Adobe says it will only be available in the latest beta. The company is also teasing more generative AI features that could be rolled out to the app this fall, giving users even more creative tools at their disposal.

This entry was posted in Presence in the News. Bookmark the permalink. Trackbacks are closed, but you can post a comment.

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

*
*

  • Find Researchers

    Use the links below to find researchers listed alphabetically by the first letter of their last name.

    A | B | C | D | E | F| G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z