Medical supervised masked autoencoder: Crafting a better masking strategy and efficient fine-tuning schedule for medical image classification

Jiawei Mao, Shujian Guo, Xuesong Yin, Yuanqi Chang, Binling Nie, Yigang Wang

Abstract: Highlights•We propose a supervised attention-driven masking strategy called SAM.•We extend SAM to fine-tuning to align pre-training and fine-tuning.•Extensive experiments show that the MSMAE exhibits superior performance.