A distribution-based adversarial attack generates quality-preserving adversarial motions for skeleton action recognition without noise perturbations, outperforming prior methods in success rate and naturalness on two datasets via a new human-aligned quality metric.
Spatio-temporal tuples transformer for skeleton-based action recognition,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Quality-Preserving Imperceptible Adversarial Attack on Skeleton-based Human Action Recognition
A distribution-based adversarial attack generates quality-preserving adversarial motions for skeleton action recognition without noise perturbations, outperforming prior methods in success rate and naturalness on two datasets via a new human-aligned quality metric.