Enhanced Approach for Classification of Ulcerative Colitis Severity in Colonoscopy Videos Using CNN

Sure Venkata Leela Lakshmi Tejaswini, Bhuvan Mittal, Jung Hwan Oh, Wallapak Tavanapong, Johnny Wong, Piet C. de Groen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations

Abstract

Ulcerative Colitis (UC) is an inflammatory bowel disease that causes inflammation, ulcers and bleeding of the colon affecting more than 500,000 people in the United States. To achieve the therapeutic objectives for UC, which are to first induce and then maintain disease remission, physicians need to evaluate the severity of UC. However, objective assessment of US is difficult because of the non-uniform nature of symptoms and large variations in disease presentation. To address this, in our previous work, we developed two different approaches in which one uses the image textures, and the other uses CNN (Convolutional Neural Network) to measure and objectively classify the severity of UC as seen on optical colonoscopy video frames. However, we found that the image texture based approach could not handle large numbers of variations in patterns, and the CNN based approach could not achieve very high accuracy. In this paper, we improve our CNN based approach in two ways to provide better accuracy for UC severity classification. We add more thorough and essential preprocessing, subdivide each class of UC severity and generate more classes for the classification to accommodate large variations in patterns. The experimental results show that the proposed preprocessing and generation of more classes can improve the overall accuracy of automated classification of the severity of UC.

Original languageEnglish (US)
Title of host publicationAdvances in Visual Computing - 14th International Symposium on Visual Computing, ISVC 2019, Proceedings
EditorsGeorge Bebis, Bahram Parvin, Richard Boyle, Darko Koracin, Daniela Ushizima, Sek Chai, Shinjiro Sueda, Xin Lin, Aidong Lu, Daniel Thalmann, Chaoli Wang, Panpan Xu
PublisherSpringer
Pages25-37
Number of pages13
ISBN (Print)9783030337223
DOIs
StatePublished - 2019
Event14th International Symposium on Visual Computing, ISVC 2019 - Nevada, United States
Duration: Oct 7 2019Oct 9 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11845 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference14th International Symposium on Visual Computing, ISVC 2019
CountryUnited States
CityNevada
Period10/7/1910/9/19

Bibliographical note

Funding Information:
Acknowledgements. This research was supported in part by a grant from the NIH (Grant #1R01DK106130-01A1). Tavanapong, Wong, and Oh have an equity interest and management role in EndoMetric Corporation, Ames, IA50014, USA, a for profit company that markets endoscopy-related software. De Groen has a financial interest in EndoMetric. The terms of this arrangement have been reviewed and approved by Iowa State University and University of Minnesota in accordance with its conflict of interest policies.

Funding Information:
This research was supported in part by a grant from the NIH (Grant #1R01DK106130-01A1). Tavanapong, Wong, and Oh have an equity interest and management role in EndoMetric Corporation, Ames, IA50014, USA, a for profit company that markets endoscopy-related software. De Groen has a financial interest in EndoMetric. The terms of this arrangement have been reviewed and approved by Iowa State University and University of Minnesota in accordance with its conflict of interest policies.

Publisher Copyright:
© 2019, Springer Nature Switzerland AG.

Copyright:
Copyright 2019 Elsevier B.V., All rights reserved.

Keywords

  • Convolutional Neural Network
  • Medical image classification
  • Medical video processing
  • Ulcerative Colitis Severity

Fingerprint Dive into the research topics of 'Enhanced Approach for Classification of Ulcerative Colitis Severity in Colonoscopy Videos Using CNN'. Together they form a unique fingerprint.

Cite this