Share Email Print
cover

Proceedings Paper

Lymphoma diagnosis in histopathology using a multi-stage visual learning approach
Author(s): Noel Codella; Mehdi Moradi; Matt Matasar; Tanveer Sveda-Mahmood; John R. Smith
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

This work evaluates the performance of a multi-stage image enhancement, segmentation, and classification approach for lymphoma recognition in hematoxylin and eosin (H and E) stained histopathology slides of excised human lymph node tissue. In the first stage, the original histology slide undergoes various image enhancement and segmentation operations, creating an additional 5 images for every slide. These new images emphasize unique aspects of the original slide, including dominant staining, staining segmentations, non-cellular groupings, and cellular groupings. For the resulting 6 total images, a collection of visual features are extracted from 3 different spatial configurations. Visual features include the first fully connected layer (4096 dimensions) of the Caffe convolutional neural network trained from ImageNet data. In total, over 200 resultant visual descriptors are extracted for each slide. Non-linear SVMs are trained over each of the over 200 descriptors, which are then input to a forward stepwise ensemble selection that optimizes a late fusion sum of logistically normalized model outputs using local hill climbing. The approach is evaluated on a public NIH dataset containing 374 images representing 3 lymphoma conditions: chronic lymphocytic leukemia (CLL), follicular lymphoma (FL), and mantle cell lymphoma (MCL). Results demonstrate a 38.4% reduction in residual error over the current state-of-art on this dataset.

Paper Details

Date Published: 23 March 2016
PDF: 7 pages
Proc. SPIE 9791, Medical Imaging 2016: Digital Pathology, 97910H (23 March 2016); doi: 10.1117/12.2217158
Show Author Affiliations
Noel Codella, IBM Thomas J. Watson Research Ctr. (United States)
Mehdi Moradi, IBM Research - Almaden (United States)
Matt Matasar, Memorial Sloan-Kettering Cancer Ctr. (United States)
Tanveer Sveda-Mahmood, IBM Research - Almaden (United States)
John R. Smith, IBM Thomas J. Watson Research Ctr. (United States)


Published in SPIE Proceedings Vol. 9791:
Medical Imaging 2016: Digital Pathology
Metin N. Gurcan; Anant Madabhushi, Editor(s)

© SPIE. Terms of Use
Back to Top