S-JEPA: A Joint Embedding Predictive Architecture for Skeletal Action Recognition.

Mohamed Abdelfattah, Alexandre Alahi

Browse the full ECCV paper archive.