ComfyUI and control net #26
Description
It might be a good idea to grab from the movies not only the pictures of the characters, but also poses and face expressions of them. Also, it might be interesting to train the model, it the future, that predicts the changes of face/position of the character from the dialog (currently, it can always be done with LLM, but probably with far less capture of the character's behaviour)
To do this, the following workflows are needed:
- Extract the ControlNet skeletal model for face or position from the image
- Apply the model to the character
These workflows need to be implemented as BrainBox ComfyUI workflows