AI RESEARCH
From Semantics to Pixels: Coarse-to-Fine Masked Autoencoders for Hierarchical Visual Understanding
arXiv CS.LG
•
ArXi:2603.09955v1 Announce Type: cross Self-supervised visual pre-