eduzhai > Applied Sciences > Engineering >

Uncertainty Quantification using Variational Inference for Biomedical Image Segmentation

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 10 pages

Abstract: Deep learning motivated by convolutional neural networks has been highlysuccessful in a range of medical imaging problems like image classification,image segmentation, image synthesis etc. However for validation andinterpretability, not only do we need the predictions made by the model butalso how confident it is while making those predictions. This is important insafety critical applications for the people to accept it. In this work, we usedan encoder decoder architecture based on variational inference techniques forsegmenting brain tumour images. We compare different backbones architectureslike U-Net, V-Net and FCN as sampling data from the conditional distributionfor the encoder. We evaluate our work on the publicly available BRATS datasetusing Dice Similarity Coefficient (DSC) and Intersection Over Union (IOU) asthe evaluation metrics. Our model outperforms previous state of the art resultswhile making use of uncertainty quantification in a principled bayesian manner.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×