Midjourney’s Unveils New ‘Vary Region’ Tool

We recently reported on Adobe’s Generative Expand feature, a new function that allows designers to extend photo canvases with AI-generated content. Now, Midjourney has unveiled its own exciting take on the generative-AI trend.

Midjourney unveiled its new “vary (region)” tool this week. The feature allows users to hone in on specific areas of an AI-generated image and randomize them to achieve desired results.

The process is simple. First, type the /imagine command into Discord and provide prompts to generate an image. Then, use the ‘vary (region) tool and highlight the area you’d like to reroll. In the example below, Midjourney has opted for a crown on a mouse.

Finally, hit submit and watch as Midjourney’s AI gets to work. Within seconds, you’ll have a selection of different variations to choose from.

Fortunately, for those of us who aren’t interested in royal mice, Midjourney allows users to generate and edit any images they generate. Unfortunately, designers can’t upload their own images for editing yet, so we’ll have to be content with the ones we generate through the program.

Midjourney notes that the feature can occasionally be unpredictable. They also specify that larger areas are better. For example, if you wanted to change a single ear on the mouse pictured above, you might run into some problems.

Despite these drawbacks, we fully expect Midjourney’s tools to improve as generative AI grows more capable. Whether Midjourney is set to steal the crown from Adobe remains to be seen, but this new feature will certainly give the giants of design some stern competition.

Max Walton

Born in Cardiff Wales, Max relocated to Brisbane when he was 12. He’s spent the last five years developing expertise in the Fintech industry. When he’s not posting about Web3, you’ll find him on a paddleboard.