Tournament Based Ranking CNN for the Cataract grading

Dohyeun Kim, Tae Joon Jun, Youngsub Eom, Cherry Kim, Daeyoung Kim

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

Abstract

Solving the classification problem, unbalanced number of dataset among the classes often causes performance degradation. Especially when some classes dominate the other classes with its large number of datasets, trained model shows low performance in identifying the dominated classes. This is common case when it comes to medical dataset. Because the case with a serious degree is not quite usual, there are imbalance in number of dataset between severe case and normal cases of diseases. Also, there is difficulty in precisely identifying grade of medical data because of vagueness between them. To solve these problems, we propose new architecture of convolutional neural network named Tournament based Ranking CNN which shows remarkable performance gain in identifying dominated classes while trading off very small accuracy loss in dominating classes. Our Approach complemented problems that occur when method of Ranking CNN that aggregates outputs of multiple binary neural network models is applied to medical data. By having tournament structure in aggregating method and using very deep pretrained binary models, our proposed model recorded 68.36% of exact match accuracy, while Ranking CNN recorded 53.40%, pretrained Resnet recorded 56.12% and CNN with linear regression recorded 57.48%. As a result, our proposed method is applied efficiently to cataract grading which have ordinal labels with imbalanced number of data among classes, also can be applied further to medical problems which have similar features to cataract and similar dataset configuration.

Original languageEnglish
Title of host publication2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1630-1636
Number of pages7
ISBN (Electronic)9781538613115
DOIs
Publication statusPublished - 2019 Jul
Event41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019 - Berlin, Germany
Duration: 2019 Jul 232019 Jul 27

Publication series

NameProceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS
ISSN (Print)1557-170X

Conference

Conference41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019
CountryGermany
CityBerlin
Period19/7/2319/7/27

Fingerprint

Cataract
Neural networks
Medical problems
Linear regression
Neural Networks (Computer)
Labels
Linear Models
Degradation
Datasets

ASJC Scopus subject areas

  • Signal Processing
  • Biomedical Engineering
  • Computer Vision and Pattern Recognition
  • Health Informatics

Cite this

Kim, D., Jun, T. J., Eom, Y., Kim, C., & Kim, D. (2019). Tournament Based Ranking CNN for the Cataract grading. In 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019 (pp. 1630-1636). [8856636] (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/EMBC.2019.8856636

Tournament Based Ranking CNN for the Cataract grading. / Kim, Dohyeun; Jun, Tae Joon; Eom, Youngsub; Kim, Cherry; Kim, Daeyoung.

2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 1630-1636 8856636 (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS).

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

Kim, D, Jun, TJ, Eom, Y, Kim, C & Kim, D 2019, Tournament Based Ranking CNN for the Cataract grading. in 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019., 8856636, Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS, Institute of Electrical and Electronics Engineers Inc., pp. 1630-1636, 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019, Berlin, Germany, 19/7/23. https://doi.org/10.1109/EMBC.2019.8856636
Kim D, Jun TJ, Eom Y, Kim C, Kim D. Tournament Based Ranking CNN for the Cataract grading. In 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 1630-1636. 8856636. (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS). https://doi.org/10.1109/EMBC.2019.8856636
Kim, Dohyeun ; Jun, Tae Joon ; Eom, Youngsub ; Kim, Cherry ; Kim, Daeyoung. / Tournament Based Ranking CNN for the Cataract grading. 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 1630-1636 (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS).
@inproceedings{7588c9d042694d26b2455ccba386fabe,
title = "Tournament Based Ranking CNN for the Cataract grading",
abstract = "Solving the classification problem, unbalanced number of dataset among the classes often causes performance degradation. Especially when some classes dominate the other classes with its large number of datasets, trained model shows low performance in identifying the dominated classes. This is common case when it comes to medical dataset. Because the case with a serious degree is not quite usual, there are imbalance in number of dataset between severe case and normal cases of diseases. Also, there is difficulty in precisely identifying grade of medical data because of vagueness between them. To solve these problems, we propose new architecture of convolutional neural network named Tournament based Ranking CNN which shows remarkable performance gain in identifying dominated classes while trading off very small accuracy loss in dominating classes. Our Approach complemented problems that occur when method of Ranking CNN that aggregates outputs of multiple binary neural network models is applied to medical data. By having tournament structure in aggregating method and using very deep pretrained binary models, our proposed model recorded 68.36{\%} of exact match accuracy, while Ranking CNN recorded 53.40{\%}, pretrained Resnet recorded 56.12{\%} and CNN with linear regression recorded 57.48{\%}. As a result, our proposed method is applied efficiently to cataract grading which have ordinal labels with imbalanced number of data among classes, also can be applied further to medical problems which have similar features to cataract and similar dataset configuration.",
author = "Dohyeun Kim and Jun, {Tae Joon} and Youngsub Eom and Cherry Kim and Daeyoung Kim",
year = "2019",
month = "7",
doi = "10.1109/EMBC.2019.8856636",
language = "English",
series = "Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "1630--1636",
booktitle = "2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019",

}

TY - GEN

T1 - Tournament Based Ranking CNN for the Cataract grading

AU - Kim, Dohyeun

AU - Jun, Tae Joon

AU - Eom, Youngsub

AU - Kim, Cherry

AU - Kim, Daeyoung

PY - 2019/7

Y1 - 2019/7

N2 - Solving the classification problem, unbalanced number of dataset among the classes often causes performance degradation. Especially when some classes dominate the other classes with its large number of datasets, trained model shows low performance in identifying the dominated classes. This is common case when it comes to medical dataset. Because the case with a serious degree is not quite usual, there are imbalance in number of dataset between severe case and normal cases of diseases. Also, there is difficulty in precisely identifying grade of medical data because of vagueness between them. To solve these problems, we propose new architecture of convolutional neural network named Tournament based Ranking CNN which shows remarkable performance gain in identifying dominated classes while trading off very small accuracy loss in dominating classes. Our Approach complemented problems that occur when method of Ranking CNN that aggregates outputs of multiple binary neural network models is applied to medical data. By having tournament structure in aggregating method and using very deep pretrained binary models, our proposed model recorded 68.36% of exact match accuracy, while Ranking CNN recorded 53.40%, pretrained Resnet recorded 56.12% and CNN with linear regression recorded 57.48%. As a result, our proposed method is applied efficiently to cataract grading which have ordinal labels with imbalanced number of data among classes, also can be applied further to medical problems which have similar features to cataract and similar dataset configuration.

AB - Solving the classification problem, unbalanced number of dataset among the classes often causes performance degradation. Especially when some classes dominate the other classes with its large number of datasets, trained model shows low performance in identifying the dominated classes. This is common case when it comes to medical dataset. Because the case with a serious degree is not quite usual, there are imbalance in number of dataset between severe case and normal cases of diseases. Also, there is difficulty in precisely identifying grade of medical data because of vagueness between them. To solve these problems, we propose new architecture of convolutional neural network named Tournament based Ranking CNN which shows remarkable performance gain in identifying dominated classes while trading off very small accuracy loss in dominating classes. Our Approach complemented problems that occur when method of Ranking CNN that aggregates outputs of multiple binary neural network models is applied to medical data. By having tournament structure in aggregating method and using very deep pretrained binary models, our proposed model recorded 68.36% of exact match accuracy, while Ranking CNN recorded 53.40%, pretrained Resnet recorded 56.12% and CNN with linear regression recorded 57.48%. As a result, our proposed method is applied efficiently to cataract grading which have ordinal labels with imbalanced number of data among classes, also can be applied further to medical problems which have similar features to cataract and similar dataset configuration.

UR - http://www.scopus.com/inward/record.url?scp=85077905848&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85077905848&partnerID=8YFLogxK

U2 - 10.1109/EMBC.2019.8856636

DO - 10.1109/EMBC.2019.8856636

M3 - Conference contribution

C2 - 31946209

AN - SCOPUS:85077905848

T3 - Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS

SP - 1630

EP - 1636

BT - 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019

PB - Institute of Electrical and Electronics Engineers Inc.

ER -