Stable Diffusion 2.0 is out and it comes with the ability to synthesize images conditioned on prompts and depth maps making it easier to maintain consistency between images of the same subject https://github.com/Stability-AI/stablediffusion#depth-conditional-stable-diffusion