Skip to content
AI & Machine Learning

ControlNet

A neural network architecture that adds spatial conditioning to diffusion models, enabling precise control over generated image structure.

ControlNet allows users to guide AI image generation with structural references like edge maps, depth maps, pose skeletons, or segmentation masks. Instead of relying solely on text prompts for composition, ControlNet takes a conditioning image that defines the spatial layout while the text prompt controls style and content. A pose skeleton ensures a generated character matches a specific body position. A depth map preserves the three-dimensional layout of a scene. An edge map maintains specific structural details. This combination of structural precision with generative flexibility enables professional-quality results that meet specific compositional requirements rather than producing random compositions from text alone.

Related Tools