We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

eess.IV

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Image and Video Processing

Title: Latent Correlation Representation Learning for Brain Tumor Segmentation with Missing MRI Modalities

Abstract: Magnetic Resonance Imaging (MRI) is a widely used imaging technique to assess brain tumor. Accurately segmenting brain tumor from MR images is the key to clinical diagnostics and treatment planning. In addition, multi-modal MR images can provide complementary information for accurate brain tumor segmentation. However, it's common to miss some imaging modalities in clinical practice. In this paper, we present a novel brain tumor segmentation algorithm with missing modalities. Since it exists a strong correlation between multi-modalities, a correlation model is proposed to specially represent the latent multi-source correlation. Thanks to the obtained correlation representation, the segmentation becomes more robust in the case of missing modality. First, the individual representation produced by each encoder is used to estimate the modality independent parameter. Then, the correlation model transforms all the individual representations to the latent multi-source correlation representations. Finally, the correlation representations across modalities are fused via attention mechanism into a shared representation to emphasize the most important features for segmentation. We evaluate our model on BraTS 2018 and BraTS 2019 dataset, it outperforms the current state-of-the-art methods and produces robust results when one or more modalities are missing.
Comments: 12 pages, 10 figures, accepted by IEEE Transactions on Image Processing (8 April 2021). arXiv admin note: text overlap with arXiv:2003.08870, arXiv:2102.03111
Subjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
Journal reference: IEEE Transactions on Image Processing On page(s): 4263-4274 Print ISSN: 1057-7149 Online ISSN: 1941-0042
DOI: 10.1109/TIP.2021.3070752
Cite as: arXiv:2104.06231 [eess.IV]
  (or arXiv:2104.06231v2 [eess.IV] for this version)

Submission history

From: Tongxue Zhou [view email]
[v1] Tue, 13 Apr 2021 14:21:09 GMT (61959kb,D)
[v2] Tue, 20 Apr 2021 13:51:09 GMT (63100kb,D)

Link back to: arXiv, form interface, contact.