The image editor lets users upload images including renderings from your computer, and then expand, crop, repaint, add, or modify things in the scene.
The image retexturing mode estimates the shape of the scene and then ‘retextures’ it so that all of the lighting, materials and surfaces are different. All image editing actions can be controlled via text prompting and region selection.
Midjourney founder David Holz says that the image editor is compatible with model personalisation, style references, character references, and image prompting.
“Along with these features, we’re testing a more nuanced and intelligent V2 AI moderator. This moderator holistically examines your prompts, images, painting masks, and output images,” he said.
“This might be the most intelligent AI moderator ever, but it’s still in early testing, and we’re trying to refine the rules it follows to be as good as possible. All of these things are very new, and we want to give the community and human moderation staff time to ease into it gently.”
Currently in an ‘early rollout and testing phase’, the new tools are available to users with Yearly memberships, those who have been subscribed for 12 months and members that have made at least 10,000 images – which will at least initially restrict use, and see users who have been waiting for such a function before adopting a subscription unable to test the functionality.
The ability to highlight features to edit – a car headlight for example – is similar to tools offered by the likes of Vizcom, which is now used widely across automotive styling studios for concepting and ideation. However, as a browser-based tool Midjourney does pose an issue around working with confidential materials – whereas Vizcom offers SOC 2 Type II compliance for data security.
