eduzhai > Applied Sciences > Engineering >

Attention based Multiple Instance Learning for Classification of Blood Cell Disorders

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 11 pages

Abstract: Red blood cells are highly deformable and present in various shapes. In bloodcell disorders, only a subset of all cells is morphologically altered andrelevant for the diagnosis. However, manually labeling of all cells islaborious, complicated and introduces inter-expert variability. We propose anattention based multiple instance learning method to classify blood samples ofpatients suffering from blood cell disorders. Cells are detected using an R-CNNarchitecture. With the features extracted for each cell, a multiple instancelearning method classifies patient samples into one out of four blood celldisorders. The attention mechanism provides a measure of the contribution ofeach cell to the overall classification and significantly improves thenetwork s classification accuracy as well as its interpretability for themedical expert.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...