The authors are experimenting with AI techniques to create game scenes, including using StableDiffusion and C

The principle and implementation of 3D scene texturing based on generative AI

The author is experimenting with AI techniques to create game scenes, including using Stable Diffusion and ControlNet to generate images. He makes reference maps in Blender and then renders depth maps, normal maps, and cue words. The authors used the SD model AyoniMix v6 and ControlNet 1.1 weights, as well as extensions to AUTOMATIC1111 and Mikubill. Finally, the author ran ControlNet using webUI and rendered the resulting image.