eduzhai > Applied Sciences > Engineering >

Interpretation of 3D CNNs for Brain MRI Data Classification

  • Save

... pages left unread,continue reading

Document pages: 13 pages

Abstract: Deep learning shows high potential for many medical image analysis tasks.Neural networks can work with full-size data without extensive preprocessingand feature generation and, thus, information loss. Recent work has shown thatthe morphological difference in specific brain regions can be found on MRI withthe means of Convolution Neural Networks (CNN). However, interpretation of theexisting models is based on a region of interest and can not be extended tovoxel-wise image interpretation on a whole image. In the current work, weconsider the classification task on a large-scale open-source dataset of younghealthy subjects -- an exploration of brain differences between men and women.In this paper, we extend the previous findings in gender differences fromdiffusion-tensor imaging on T1 brain MRI scans. We provide the voxel-wise 3DCNN interpretation comparing the results of three interpretation methods:Meaningful Perturbations, Grad CAM and Guided Backpropagation, and contributewith the open-source library.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×