San Francisco-based AI research lab Midjourney has gained immense popularity in recent months due to its outstanding text-to-image feature and now there is upgraded news that it has introduced a unique feature called “Character Reference.” The just launched new feature basically aims to address a common struggle usually faced by users while generating consistent characters across multiple reference images.
Let us find out how the new Character Reference feature works on the platform
Users simply need to type “–cref URL” to initiate after their prompt along with the character image URL. Once done, the reference ‘strength’ can be adjusted using “–cw” from 100 to 0. A strength of 100 (default) considers various aspects like face, hair and clothes, while a strength of 0 focuses solely on the face, ideal for alterations in outfits or hairstyles. Blending of information from multiple images can be achieved by using “–cref URL1 URL2.” It is similar to the multiple image or style prompts.
Generating Character Image
Users can retrieve or generate the character image URL through the platform. They can use “–cref” and thereafter type the URL to create the character in various settings.
Adjusting Image Variance
Control over how closely the new image reproduces the original character is facilitated. One can use “–cw” followed by a number (ranging from 1 to 100) to adjust the variance. Do note that lower numbers offer more variation and higher numbers closely replicate the original reference.
Advanced Options and Considerations
Multiple “–cref” tags with respective URLs can be used to blend information from multiple images. In the web alpha version of Midjourney one can have the convenience to drag or paste an image, selecting it as a prompt, style reference, or character reference.
The new feature has hit headlines as well as social media. It has garnered positive feedback from users. People are appreciating its utility and potential for creating consistent and diverse characters effortlessly.
