X-Invariant Contrastive Augmentation and Representation Learning for Semi-Supervised Skeleton-Based Action Recognition | IEEE Journals & Magazine | IEEE Xplore