Elevating Skeleton-Based Action Recognition with Efficient Multi-Modality Self-Supervision

Open in new window