eduzhai > Applied Sciences > Engineering >

Visualization for Histopathology Images using Graph Convolutional Neural Networks

  • Save

... pages left unread,continue reading

Document pages: 5 pages

Abstract: With the increase in the use of deep learning for computer-aided diagnosis inmedical images, the criticism of the black-box nature of the deep learningmodels is also on the rise. The medical community needs interpretable modelsfor both due diligence and advancing the understanding of disease and treatmentmechanisms. In histology, in particular, while there is rich detail availableat the cellular level and that of spatial relationships between cells, it isdifficult to modify convolutional neural networks to point out the relevantvisual features. We adopt an approach to model histology tissue as a graph ofnuclei and develop a graph convolutional network framework based on attentionmechanism and node occlusion for disease diagnosis. The proposed methodhighlights the relative contribution of each cell nucleus in the whole-slideimage. Our visualization of such networks trained to distinguish betweeninvasive and in-situ breast cancers, and Gleason 3 and 4 prostate cancersgenerate interpretable visual maps that correspond well with our understandingof the structures that are important to experts for their diagnosis.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...