Glimpse Clouds: Human Activity Recognition from Unstructured Feature Points

CVPR 2018 Fabien BaradelChristian WolfJulien MilleGraham W. Taylor

We propose a method for human activity recognition from RGB data that does not rely on any pose information during test time and does not explicitly calculate pose information internally. Instead, a visual attention module learns to predict glimpse sequences in each frame... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Action Recognition In Videos NTU RGB+D Glimpse Clouds (RGB only) Accuracy (CS) 86.6 # 4
Action Recognition In Videos NTU RGB+D Glimpse Clouds (RGB only) Accuracy (CV) 93.2 # 3
Skeleton Based Action Recognition N-UCLA Glimpse Clouds Accuracy 87.6% # 5