Classification of hematologic malignancies using texton signatures

  • Authors:

  • Oncel Tuzel;Lin Yang;Peter Meer;David J. Foran

  • Affiliations:

  • Rutgers University, Department of Computer Science, 08854, Piscataway, NJ, USA and CAIP Center, 96 Frelinghuysen Road, 08854, Piscataway, NJ, USA;Rutgers University, Department of Electrical and Computer Engineering, 08854, Piscataway, NJ, USA and University of Medicine and Dentistry of New Jersey, Center for Biomedical Imaging and Informat ...;Rutgers University, Department of Computer Science, 08854, Piscataway, NJ, USA and Rutgers University, Department of Electrical and Computer Engineering, 08854, Piscataway, NJ, USA;University of Medicine and Dentistry of New Jersey, Center for Biomedical Imaging and Informatics, 08854, Piscataway, NJ, USA

  • Venue:
  • Pattern Analysis & Applications
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

We describe a decision support system to distinguish among hematology cases directly from microscopic specimens. The system uses an image database containing digitized specimens from normal and four different hematologic malignancies. Initially, the nuclei and cytoplasmic components of the specimens are segmented using a robust color gradient vector flow active contour model. Using a few cell images from each class, the basic texture elements (textons) for the nuclei and cytoplasm are learned, and the cells are represented through texton histograms. We propose to use support vector machines on the texton histogram based cell representation and achieve major improvement over the commonly used classification methods in texture research. Experiments with 3,691 cell images from 105 patients which originated from four different hospitals indicate more than 84% classification performance for individual cells and 89% for case based classification for the five class problem.