This is a simple workflow to generate elaborate visualization based on photo input. In this particular case, we take a photo of a IRL clay made sculpture as a base. This photo is run through DepthAnything to extract a legit depthmap from this photo. ControlNET Depth will use this depth information to generate a fresh reinterpretation of the input based on the prompt cloud.

*NOTE: the missing node is called DepthAnything - i have no clue, why it is not included?!? - it is saved in the JSON anyway.