New mammography datasets and generalization framework improve AI accuracy

Researchers have introduced two new datasets, BreastMammo and DenseMammo, designed to improve AI model generalization in mammography across different clinical sites. They also proposed a domain generalization framework that uses foreground-only histogram matching to address variations in image acquisition styles. This approach, when tested with a Swin Transformer backbone, achieved a 98.32% AUC for density classification and outperformed existing methods like MixStyle and discrete Fourier transform-based frameworks on external datasets. AI

IMPACT This research could lead to more reliable AI diagnostic tools for mammography, improving consistency across different healthcare providers.

RANK_REASON The cluster contains an academic paper introducing new datasets and a novel framework for domain generalization in medical imaging. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• BreastMammo
• DenseMammo
• discrete Fourier transform
• LUMINA
• MixStyle
• Swin Transformer
• TNMammo

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →