ColorMAE: Data-independent Masking for Masked AutoEncoders

Go to Project Site
Computer vision