Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks

Nima Tajbakhsh, Suryakanth R. Gurudu, Jianming Liang

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

49 Citations (Scopus)

Abstract

Computer-aided polyp detection in colonoscopy videos has been the subject of research for over the past decade. However, despite significant advances, automatic polyp detection is still an unsolved problem. In this paper, we propose a new polyp detection method based on a unique 3-way image presentation and convolutional neural networks. Our method learns a variety of polyp features such as color, texture, shape, and temporal information in multiple scales, enabling a more accurate polyp localization. Given a polyp candidate, a set of convolution neural networks - each specialized in one type of features - are applied in the vicinity of the candidate and then their results are aggregated to either accept or reject the candidate. Our experimental results based on our collection of videos, which to our knowledge is the largest annotated polyp database, shows a remarkable performance improvement over the state-of-the-art, significantly reducing the number of false positives in nearly all operating points. In addition, we propose a new performance curve, demonstrating that our new method significantly decreases polyp detection latency, which is defined as the time from the first appearance of a polyp in the video to the time of its first detection by our method.

Original languageEnglish (US)
Title of host publicationProceedings - International Symposium on Biomedical Imaging
PublisherIEEE Computer Society
Pages79-83
Number of pages5
Volume2015-July
ISBN (Print)9781479923748
DOIs
StatePublished - Jul 21 2015
Event12th IEEE International Symposium on Biomedical Imaging, ISBI 2015 - Brooklyn, United States
Duration: Apr 16 2015Apr 19 2015

Other

Other12th IEEE International Symposium on Biomedical Imaging, ISBI 2015
CountryUnited States
CityBrooklyn
Period4/16/154/19/15

Fingerprint

Colonoscopy
Polyps
Neural networks
Convolution
Textures
Color
Databases
Research

ASJC Scopus subject areas

  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging

Cite this

Tajbakhsh, N., Gurudu, S. R., & Liang, J. (2015). Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks. In Proceedings - International Symposium on Biomedical Imaging (Vol. 2015-July, pp. 79-83). [7163821] IEEE Computer Society. https://doi.org/10.1109/ISBI.2015.7163821

Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks. / Tajbakhsh, Nima; Gurudu, Suryakanth R.; Liang, Jianming.

Proceedings - International Symposium on Biomedical Imaging. Vol. 2015-July IEEE Computer Society, 2015. p. 79-83 7163821.

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

Tajbakhsh, N, Gurudu, SR & Liang, J 2015, Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks. in Proceedings - International Symposium on Biomedical Imaging. vol. 2015-July, 7163821, IEEE Computer Society, pp. 79-83, 12th IEEE International Symposium on Biomedical Imaging, ISBI 2015, Brooklyn, United States, 4/16/15. https://doi.org/10.1109/ISBI.2015.7163821
Tajbakhsh N, Gurudu SR, Liang J. Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks. In Proceedings - International Symposium on Biomedical Imaging. Vol. 2015-July. IEEE Computer Society. 2015. p. 79-83. 7163821 https://doi.org/10.1109/ISBI.2015.7163821
Tajbakhsh, Nima ; Gurudu, Suryakanth R. ; Liang, Jianming. / Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks. Proceedings - International Symposium on Biomedical Imaging. Vol. 2015-July IEEE Computer Society, 2015. pp. 79-83
@inproceedings{023384ea51f54fbda7e56cb0bf6ad1f9,
title = "Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks",
abstract = "Computer-aided polyp detection in colonoscopy videos has been the subject of research for over the past decade. However, despite significant advances, automatic polyp detection is still an unsolved problem. In this paper, we propose a new polyp detection method based on a unique 3-way image presentation and convolutional neural networks. Our method learns a variety of polyp features such as color, texture, shape, and temporal information in multiple scales, enabling a more accurate polyp localization. Given a polyp candidate, a set of convolution neural networks - each specialized in one type of features - are applied in the vicinity of the candidate and then their results are aggregated to either accept or reject the candidate. Our experimental results based on our collection of videos, which to our knowledge is the largest annotated polyp database, shows a remarkable performance improvement over the state-of-the-art, significantly reducing the number of false positives in nearly all operating points. In addition, we propose a new performance curve, demonstrating that our new method significantly decreases polyp detection latency, which is defined as the time from the first appearance of a polyp in the video to the time of its first detection by our method.",
author = "Nima Tajbakhsh and Gurudu, {Suryakanth R.} and Jianming Liang",
year = "2015",
month = "7",
day = "21",
doi = "10.1109/ISBI.2015.7163821",
language = "English (US)",
isbn = "9781479923748",
volume = "2015-July",
pages = "79--83",
booktitle = "Proceedings - International Symposium on Biomedical Imaging",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks

AU - Tajbakhsh, Nima

AU - Gurudu, Suryakanth R.

AU - Liang, Jianming

PY - 2015/7/21

Y1 - 2015/7/21

N2 - Computer-aided polyp detection in colonoscopy videos has been the subject of research for over the past decade. However, despite significant advances, automatic polyp detection is still an unsolved problem. In this paper, we propose a new polyp detection method based on a unique 3-way image presentation and convolutional neural networks. Our method learns a variety of polyp features such as color, texture, shape, and temporal information in multiple scales, enabling a more accurate polyp localization. Given a polyp candidate, a set of convolution neural networks - each specialized in one type of features - are applied in the vicinity of the candidate and then their results are aggregated to either accept or reject the candidate. Our experimental results based on our collection of videos, which to our knowledge is the largest annotated polyp database, shows a remarkable performance improvement over the state-of-the-art, significantly reducing the number of false positives in nearly all operating points. In addition, we propose a new performance curve, demonstrating that our new method significantly decreases polyp detection latency, which is defined as the time from the first appearance of a polyp in the video to the time of its first detection by our method.

AB - Computer-aided polyp detection in colonoscopy videos has been the subject of research for over the past decade. However, despite significant advances, automatic polyp detection is still an unsolved problem. In this paper, we propose a new polyp detection method based on a unique 3-way image presentation and convolutional neural networks. Our method learns a variety of polyp features such as color, texture, shape, and temporal information in multiple scales, enabling a more accurate polyp localization. Given a polyp candidate, a set of convolution neural networks - each specialized in one type of features - are applied in the vicinity of the candidate and then their results are aggregated to either accept or reject the candidate. Our experimental results based on our collection of videos, which to our knowledge is the largest annotated polyp database, shows a remarkable performance improvement over the state-of-the-art, significantly reducing the number of false positives in nearly all operating points. In addition, we propose a new performance curve, demonstrating that our new method significantly decreases polyp detection latency, which is defined as the time from the first appearance of a polyp in the video to the time of its first detection by our method.

UR - http://www.scopus.com/inward/record.url?scp=84944325843&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84944325843&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2015.7163821

DO - 10.1109/ISBI.2015.7163821

M3 - Conference contribution

SN - 9781479923748

VL - 2015-July

SP - 79

EP - 83

BT - Proceedings - International Symposium on Biomedical Imaging

PB - IEEE Computer Society

ER -