Reproduce “Deblurring Masked Autoencoder is Better Recipe for Ultrasound Image Recognition” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Qingbo Kang, Jun Gao, Kang Li, Qicheng Lao
The deblurring masked autoencoder (MAE) adapts self-supervised pretraining to ultrasound by blurring images, masking patches, and reconstructing the original unblurred image. During transfer, it also blurs classifier inputs to reduce distribution shift. On the West China Hospital thyroid-nodule classification experiment, it reports F1 of 88.48±0.50%, versus 87.54±0.62% for ultrasound-pretrained vanilla MAE; evidence is limited to this task and comparison.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.