TM2T: Stochastic and Tokenized Modeling for the Reciprocal Generation of 3D Human Motions and Texts.

Chuan Guo Xinxin Xuo Sen Wang Li Cheng

Published in: CoRR (2022)

Keyphrases

human motion
stochastic models
motion sequences
motion capture
text generation
image sequences
video sequences
multi modal
human motion tracking
neural network
linear dynamical systems
gait recognition
body parts
spatio temporal
keywords
computer vision
machine learning