Social-MAE: Social Masked Autoencoder for Multi-Person Motion Representation Learning.

Mahsa Ehsanpour, Ian Reid, Hamid Rezatofighi

Browse the full ICRA paper archive.