1 / 13

In Silico Brain Tumor Research Center Emory University, Atlanta, GA

Classification of Brain Tumor Regions S. Cholleti, *L. Cooper, J. Kong, C. Chisolm, D. Brat, D. Gutman, T. Pan, A. Sharma, C. Moreno, T. Kurc, J. Saltz. In Silico Brain Tumor Research Center Emory University, Atlanta, GA. In Silico Brain Tumor Research. molecular. neuroimaging. Integrated

Télécharger la présentation

In Silico Brain Tumor Research Center Emory University, Atlanta, GA

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Classification of Brain Tumor RegionsS. Cholleti, *L. Cooper, J. Kong, C. Chisolm, D. Brat, D. Gutman, T. Pan, A. Sharma, C. Moreno, T. Kurc, J. Saltz In Silico Brain Tumor Research Center Emory University, Atlanta, GA

  2. In Silico Brain Tumor Research molecular neuroimaging Integrated Analysis histology clincal\pathology Datasets: In Silico Research Centers of Excellence

  3. Morphometry of the Gliomas Nuclear Morphology: Oligodendroglioma Astrocytoma Vessel Morphology: Necrosis:

  4. Morphometric Analysis Scientific Queries PAIS Database Parallel Matlab ? (90+ Million Nuclei)

  5. Morphological Correlates of Genomic Analysis Nuclear Classification ? Nuclear Characterization Classical Region Filtering Proneural Nuclear priors Mesenchymal (Neoplastic Oligodendroglia, Neoplastic Astrocytes, Reactive Endothelial, ...) Class Summary Statistics Neural

  6. Morphological Correlates of Genomic Analysis ? Nuclear Characterization Nuclear Classification Tissue Classification Nuclear Priors Classical Proneural Mesenchymal Neural (Neoplastic Oligodendroglia, Neoplastic Astrocytes, Reactive Endothelial, ...) Class Summary Statistics

  7. Region Classification • Classify regions as normal or tumor • exclude nuclei in normal tissue regions • conditional probabilities for nuclear classification • texton approach • Multiple layers of classification add robustness • Combines supervised and unsupervised classifiers • References • Malik, J., Belongie, S., Shi, J., and Leung, T. 1999. Textons, contours and regions: Cue integration in image segmentation. In Proceedings IEEE 7th International Conference on Computer Vision, Corfu, Greece, pp. 918–925. • O. Tuzel, L. Yang, P. Meer, and D. J. Foran. Classification of hematologic malignancies using texton signatures. Pattern Anal. Appl., 10(4):277-290,2007. • M. Varma and A. Zisserman. Texture classification: Are filter banks necessary? In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 691-698, 2003.

  8. Tissue Classifier: Training For each class (texture classification): Training Regions Extract “Textures” Texton Library For each training region: Train Region Classifier Region “Textures” Texton Histogram SVM

  9. Tissue Classifier: Testing Test Region Texton Library Region Classification Region “Textures” Texton Histogram SVM

  10. Dataset • Human Annotated regions • 18 whole-slide images • Normal, GBM (IV), Astrocytoma (II & III), Oligodendroglioma (II & III), Oligoastrocytoma (II & III)

  11. Experiment and Results • 30 x 2 cross-validation • Randomly pick 50% data for training and 50% for testing. • Classification accuracy: • Average(correct regions / total regions)

  12. Extension: Region Masking

  13. Questions

More Related