LEAD: Latent Realignment for Human Motion Diffusion
Sign Up to like & getrecommendations! Published in 2024 at "Computer Graphics Forum"
DOI: 10.1111/cgf.70093
Abstract: Our goal is to generate realistic human motion from natural language. Modern methods often face a trade‐off between model expressiveness and text‐to‐motion (T2M) alignment. Some align text and motion latent spaces but sacrifice expressiveness; others… read more here.
Keywords: text motion; human motion; realignment; language ... See more keywords