Midjourney: New external image editor and image retexturing mode
A quick overview of Midjourney's latest features
On October 24th, Midjourney released a new external image editor and image retexturing mode, allowing users to upload images for inpainting, outpainting, or modifying them.
The new external image editor differs from the existing Web Editor in that it allows you to edit images created both outside of and within Midjourney.
The image retexturing mode modifies all of the lighting, materials, and surfaces while maintaining the position of the subjects or objects.
Both the image editor and the image retexturing mode support model personalization, style references, character references, and image prompting.
In short, the new features can blur the line between reality and imagination by allowing users to edit photographs taken with their cameras and create new photorealistic images that are difficult to tell if they were created using AI.
Let's look at some use cases for these new features.
The basic functions
I added a sketched illustration of a man to the edit menu. The illustration is a piece of vintage art sourced from a public gallery archive.
I retextured the image with a simple prompt: "cinestill photograph of a cyborg thinking in the style of sci-fi movie
," and I was pleased with how easily the sketch has been transformed into a photorealistic image.
There are two methods for exporting the image:
"Upscale to Gallery" will upscale the image and display it in the Midjourney website's standard "Organize" menu.
"Download Image" to download and save the file
One interesting new feature is the "Suggest Prompt" button, which creates a prompt similar to Discord's /describe
, but with only one prompt rather than the usual four.
The image was edited to remove the border and correct the abnormal hand. Instead of addressing each issue separately, we can delete them and run Submit to fix these issues all at once. The image is also resized to a 16:9 aspect ratio.
When the prompt was unchanged, the edit produced this result.
This was the result after using the suggested prompt to edit the image and then retextured.
The retexturing mode, guided by the Suggest Prompt, altered the gender of the main subject and other details, resulting in an even more polished result than before.
The effects of Character (Cref) and Style References (Sref)
This character was used to generate a consistent character.
This style was used to create a consistent style.
The result of character consistency using the keywords "portrait photography, documentary."
The result of character consistency using the keywords "portrait, fine art."
The result of using the style reference alone.
The effects of combining style and character references.
The retextured mode can be used to pose a reference character with --cref while maintaining style consistency with --sref.
The new features broaden the use cases of Sref by allowing users to easily test the effects of various Sref codes on an image in order to select the best Sref code to style it.
The new features also allows for the simple transformation of illustration-type images to photographic images and vice versa while keeping the subject's position unchanged.
Recent articles
I hope you liked this article!
Please subscribe, like, share, and comment so that more people can discover Geeky Curiosity newsletter.
Geeky Curiosity is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.