Disentangling Subject-Dependent/-Independent Representations for 2D Motion Retargeting.

Fanglu Xie Go Irie Tatsushi Matsubayashi

Published in: ICASSP (2021)

Keyphrases

motion estimation
completely independent
video sequences
higher level
feature points
statistically independent
image motion
neural network
optical flow
image retrieval
three dimensional
image features
image content
motion capture
image sequences
distributed representations
image retargeting
information retrieval