REALM: A Coarse-to-Fine Generative Framework for Embodied Reactive Listening
REALM generates reactive listener facial motion with coarse trajectories plus stochastic expression residuals, including Ameca deployment.
1 stories, newest first · one card per story
REALM generates reactive listener facial motion with coarse trajectories plus stochastic expression residuals, including Ameca deployment.
Today 1.6K readers · 2.5K views · 779 reads · all time 189K views