Midjourney released the newest version of its AI-powered image generator service last week, which notably includes a new ‘zoom out’ feature that expands the borders of an existing AI-generated image, similarly to zooming out with a camera lens. However, at the moment this feature only works on images generated within Midjourney, for everyone else, you’ll still need to use standard photo editing software.
The feature appears as a button underneath generated images and users have two options: ‘Zoom Out 1.5x’ and ‘Zoom Out 2x.’ There are also options to turn a non-square image into a square one and a ‘Custom Zoom’ button which allows users to edit text prompts as well as the aspect ratio.
This comes as part of a recent series of updates since March 2022 which have aimed to improve realism and image detail, resulting in the creation of realistic images of Pope Francis and Donald Trump that raised concern about deepfakes spreading on social media. Unlike Adobe with its own AI-powered image generator Firefly, Midjourney has not officially revealed the exact contents of its training data.
Additional updates to Midjourney v5.2 include an improved aesthetic system that promises better image quality and a stronger ‘stylise’ command that influences how non-realistic an image looks. There is also a new shorten command for seeing what words in a text prompt are affecting the image and which ones aren’t.
- Is Photoshop Generative Fill good or bad for photographers?
- With Generative Fill, Adobe Photoshop takes AI to a whole new level
- Firefly for Adobe Illustrator announced: Adobe’s AI editing tools showcased at major London event
- Will Adobe’s AI fixation ruin photography as we know it?