1. MidJourney has a new character consistency feature that works by using a reference image for specific character traits found in the source image.
2. The feature appends a code to the end of the prompt and works best with MidJourney images, allowing for multiple characters in a scene by using multiple source images.
3. The main use cases include creating graphic novels, short video clips, and storytelling across various mediums, making it easier for writers and filmmakers to visualize their creations.
MidJourney has introduced a new character consistency feature in their AI image generator, allowing users to maintain specific character traits across images and video. By appending a code at the end of the prompt and using MidJourney images as a source, users can control how closely the generated images stick to the original character.
This feature is especially useful for creating graphic novels or short video clips with consistent characters. By adjusting the parameters, users can control how closely the generated images resemble the source image’s traits, such as clothing, facial features, and hairstyle. Multiple characters can be added to a scene by including multiple reference images in the prompt.
In a test using a character named Jack Sloane, an old-school detective, the character consistency feature successfully maintained traits like his weathered face and scowl across different generations of images. Users can adjust the parameters to create variations like changing outfits or adding additional characters to a scene.
While the feature worked well for maintaining character consistency, there were some limitations when adding multiple characters to a scene. Overall, MidJourney’s character consistency feature opens up new possibilities for content creation in various mediums, allowing for visualizing creations more easily and turning ideas into reality with generated storytelling.