I want to combine USO (on Flux-1-Dev) with ControlNet so I can keep the exact identity/shape of my subject (e.g. a specific car), but also force the correct perspective/pose so it sits on a road in a target landscape. I also want to change the background as little as possible.
So USO would be used for subject (car) and controlnet canny/depth/union for background image.
Could you please share a minimal ComfyUI JSON (or diagram) for USO + ControlNet Depth/Canny that keeps identity, aligns perspective, and leaves the background mostly intact?