Bayesian Graph Convolution LSTM for Skeleton Based Action Recognition

ICCV 2019 Rui Zhao Kang Wang Hui Su Qiang Ji

We propose a framework for recognizing human actions from skeleton data by modeling the underlying dynamic process that generates the motion pattern. We capture three major factors that contribute to the complexity of the motion pattern including spatial dependencies among body joints, temporal dependencies of body poses, and variation among subjects in action execution... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK LEADERBOARD
Skeleton Based Action Recognition NTU RGB+D Bayesian GC-LSTM Accuracy (CV) 89.0 # 34
Accuracy (CS) 81.8 # 36