Object-Centric Video Representation for Long-Term Action Anticipation | ComputerVisionFoundation Videos | Podwise