Soft Labeling by Distilling Anatomical Knowledge for Improved MS Lesion Segmentation

@article{Kats2019SoftLB,
  title={Soft Labeling by Distilling Anatomical Knowledge for Improved MS Lesion Segmentation},
  author={Eytan Kats and J. Goldberger and H. Greenspan},
  journal={2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019)},
  year={2019},
  pages={1563-1566}
}
  • Eytan Kats, J. Goldberger, H. Greenspan
  • Published 2019
  • Computer Science
  • 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019)
  • This paper explores the use of a soft ground-truth mask (“soft mask”) to train a Fully Convolutional Neural Network (FCNN) for segmentation of Multiple Sclerosis (MS) lesions. Detection and segmentation of MS lesions is a complex task largely due to the extreme unbalanced data, with very small number of lesion pixels that can be used for training. Utilizing the anatomical knowledge that the lesion surrounding pixels may also include some lesion level information, we suggest to increase the data… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    A Soft STAPLE Algorithm Combined with Anatomical Knowledge
    3
    Learning Probabilistic Fusion of Multilabel Lesion Contours
    Unpaired Multi-Modal Segmentation via Knowledge Distillation
    12
    Accurate 2D soft segmentation of medical image via SoftGAN network
    Superpixel-Guided Label Softening for Medical Image Segmentation

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 13 REFERENCES
    Improving automated multiple sclerosis lesion segmentation with a cascaded 3D convolutional neural network approach
    140
    Patch-Based Segmentation with Spatial Consistency: Application to MS Lesions in Brain MRI
    30
    U-Net: Convolutional Networks for Biomedical Image Segmentation
    14431
    Brain tumor segmentation with Deep Neural Networks
    1219
    Longitudinal multiple sclerosis lesion segmentation: Resource and challenge
    105