Improving Pose Estimation for Generative AI
submited by
Style Pass
Turning human poses into skeletal stick-figures and back into new images via generative AI is fraught with pitfalls, particularly when the pose in question is uncommon, taken from an unusual angle, or in some way ‘out of distribution’ for what the target generative system is expecting. Among systems that perform these tasks for Stable Diffusion, ControlNet’s openpose module has become very popular in the last year or so – but new research has improved on it, bringing us nearer to the dream of purely generative video generation.
Read more blog.metaphy...
