Let’s go for a jog
Jogging is better with a friend!
A friend asked me how to do pose control inside ComfyUI. Unlike Automatic1111, it’s slightly more involved to do control net inside ComfyUI because you will have to wire everything together yourself. But once that is done, it offers a lot more flexibility in how you would flow the different parts.
Using pose control allows you to take the poses of an image but change everything else. Here I use it to translate an illustration I made in Midjourney and place the figures inside different environments. In Stable Diffusion, faces can be kept relatively stable with the same prompt, so minimal fuss is required to force consistency.
There are techniques you can use to change the faces for each render, but since I do prefer to have the characters stay relatively stable, this becomes a good to have.
I made these as a technical demo to show how to do pose control with ComfyUI. I will post a step by step tutorial in the future.
Stable Diffusion, Virile Fusion v3 Beta 1, OpenPose control net with DWPose, SD Ultimate Upscale.
