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-