Midjourney, a leader in neural network technology for generating images from text descriptions, has announced the testing of new features. These include an external image editor, a texture editing mode, and a sophisticated AI-based moderation system.
New Editor and Texture Transformation Features
The new AI-powered image editor allows users to upload photos from their devices and modify them. Users can expand, crop, redraw, or insert objects into the scene using text-based prompts or by selecting areas in the image. The editor also supports personalizing AI models, using reference characters, and generating automatic prompts based on provided images.
Another noteworthy addition is the texture editing mode, which analyzes the shape and structure of a scene to adjust lighting, materials, and textures. This feature makes it possible to create entirely new visual effects without altering the core composition of the original image, giving users more flexibility in creative projects.
AI Moderation System and Availability
Midjourney has also rolled out a second-generation AI-based moderation system (V2). This advanced system monitors not only text inputs but also masks, generated images, and outputs. According to the developers, this is the most refined AI moderator available today. “We hope that some of its features will be incorporated into standard generation pipelines in the future to reduce false positives and provide users with greater creative freedom,” they said.
These new features are currently available to users who meet certain criteria—those who have generated over 10,000 images, hold an annual subscription, or have maintained a monthly subscription for the past 12 months, adds NIXSOLUTIONS.
The developers acknowledge a few minor issues, such as challenges when working with small areas in images, and request patience from users while improvements are made. We’ll keep you updated as these features evolve and become more accessible.