Leaders in Boston will have the opportunity to engage in an exclusive night of networking, insights, and conversation on March 27. This event promises to bring together industry professionals for a valuable exchange of ideas and perspectives.
Meanwhile, Midjourney, a prominent AI image generating service, has unveiled a new feature that is set to revolutionize the way users interact with the platform. This innovative addition allows users to consistently recreate characters across different images, overcoming a common challenge faced by AI image generators.
Traditionally, AI image generators rely on diffusion models to create new content for every prompt entered. However, the new “Character Reference” feature on Midjourney introduces a level of consistency that is particularly beneficial for generative AI imagery, especially when crafting visual narratives that require continuity.
To utilize this feature, users can simply add the tag “-cref” to their prompts in the Midjourney Discord and paste a URL of a previously generated character. This enables users to match facial features, body type, and clothing in order to maintain character consistency. Additionally, users have the ability to adjust the level of similarity to the original character through a variable called “cw.”
Artists and creators are already experimenting with the feature and providing valuable feedback to enhance its functionality. It is important to note that the “Character Reference” feature is specifically designed for characters created from Midjourney images and is not intended for real people or photos.
Furthermore, advanced features allow users to merge multiple characters into one image by using multiple URLs. The feature is currently available in the web alpha version of Midjourney V6 and is subject to change as the platform continues to evolve. The official beta version of Midjourney V6 is on the horizon, and feedback on the new feature is encouraged to further improve user experience.
