• Train an attention network using skeleton data from Motion Capture.

    • Even if the number of joints differs, a common latent space is created, which can be used for retargeting body motion.
  • Is it possible to translate movements from a ten-headed body to a two-headed body?