ControlNet is a neural network architecture that adds spatial conditioning control to diffusion model image generation, allowing architects to use line drawings, depth maps, edge detection outputs and other structural guides to constrain AI-generated imagery. Without ControlNet, diffusion models generate images from prompts but have no direct way to enforce spatial layout, depth or edge structure. ControlNet solves this by taking an additional input image and using it as a conditioning signal, giving architects the ability to generate AI imagery that follows their design geometry rather than producing arbitrary compositions.
.jpg)
