GFP predicts hierarchical high-level features from a jointly trained lightweight target network instead of reconstructing joint coordinates, speeding up masked skeleton pretraining 6.2x while improving downstream accuracy.
S-jepa: A joint embedding predictive architecture for skeletal action recog- nition
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
baseline 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
baseline 1polarities
baseline 1representative citing papers
citing papers explorer
-
Towards Efficient General Feature Prediction in Masked Skeleton Modeling
GFP predicts hierarchical high-level features from a jointly trained lightweight target network instead of reconstructing joint coordinates, speeding up masked skeleton pretraining 6.2x while improving downstream accuracy.