eduzhai > Applied Sciences > Engineering >

Federated Learning for Breast Density Classification A Real-World Implementation

  • king
  • (0) Download
  • 20210507
  • Save

... pages left unread,continue reading

Document pages: 11 pages

Abstract: Building robust deep learning-based models requires large quantities ofdiverse training data. In this study, we investigate the use of federatedlearning (FL) to build medical imaging classification models in a real-worldcollaborative setting. Seven clinical institutions from across the world joinedthis FL effort to train a model for breast density classification based onBreast Imaging, Reporting & Data System (BI-RADS). We show that despitesubstantial differences among the datasets from all sites (mammography system,class distribution, and data set size) and without centralizing data, we cansuccessfully train AI models in federation. The results show that modelstrained using FL perform 6.3 on average better than their counterparts trainedon an institute s local data alone. Furthermore, we show a 45.8 relativeimprovement in the models generalizability when evaluated on the otherparticipating sites testing data.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×