Automatic detection of early esophageal cancer with CNNS using transfer learning

Sjors van Riel, Fons van der Sommen, Sveta Zinger, Erik J. Schoon, Peter H.N. de With

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

3 Citations (Scopus)
2 Downloads (Pure)

Abstract

The incidence of Esophageal Adenocarcinoma (EAC), a form of esophageal cancer, has rapidly increased in recent years. Dysplastic tissue can be removed endoscopically at an early stage, and since survival chances of patients are limited at later stages of the disease, early detection is of key impor- tance. Recently, several CAD systems for HD endoscopic images have been proposed, but these are computationally expensive, making them unfit for clinical use requiring real- time analysis. In this paper, we present a novel approach for early esophageal cancer detection using Transfer Learning with CNNs. Given the small amount of annotated data, CNN Codes are applied, where intermediate layers of the net- work are used as features for conventional classifiers. Various classifiers are combined with four of the most widely-used networks. Additionally, sliding windows are used to obtain a coarse-grained annotation indicating any possible cancerous regions. This approach outperforms the current state-of-the-art with a frame-based AUC of 0.92, while allowing both near real-time prediction and annotation at 2 fps, in a MATLAB-based framework.
Original languageEnglish
Title of host publicationProc. IEEE International Conference on Image Processing (ICIP)
Place of PublicationPiscataway
PublisherIEEE Computer Society
Pages1383-1387
Number of pages5
ISBN (Electronic)978-1-4799-7061-2
ISBN (Print)978-1-4799-7062-9
DOIs
Publication statusPublished - Oct 2018
Event25th IEEE International Conference on Image Processing (ICIP 2018) - Megaron Athens International Conference Centre, Athens, Greece
Duration: 7 Oct 201810 Oct 2018
Conference number: 25
http://athenscvb.gr/en/content/25-international-conference-image-processing-icip-2018

Conference

Conference25th IEEE International Conference on Image Processing (ICIP 2018)
Abbreviated titleICIP 2018
CountryGreece
CityAthens
Period7/10/1810/10/18
Internet address

Fingerprint

Classifiers
MATLAB
Computer aided design
Tissue

Cite this

van Riel, S., van der Sommen, F., Zinger, S., Schoon, E. J., & de With, P. H. N. (2018). Automatic detection of early esophageal cancer with CNNS using transfer learning. In Proc. IEEE International Conference on Image Processing (ICIP) (pp. 1383-1387). [8451771] Piscataway: IEEE Computer Society. https://doi.org/10.1109/ICIP.2018.8451771
van Riel, Sjors ; van der Sommen, Fons ; Zinger, Sveta ; Schoon, Erik J. ; de With, Peter H.N. / Automatic detection of early esophageal cancer with CNNS using transfer learning. Proc. IEEE International Conference on Image Processing (ICIP). Piscataway : IEEE Computer Society, 2018. pp. 1383-1387
@inproceedings{d125bea34e38422c8bba4334424a063d,
title = "Automatic detection of early esophageal cancer with CNNS using transfer learning",
abstract = "The incidence of Esophageal Adenocarcinoma (EAC), a form of esophageal cancer, has rapidly increased in recent years. Dysplastic tissue can be removed endoscopically at an early stage, and since survival chances of patients are limited at later stages of the disease, early detection is of key impor- tance. Recently, several CAD systems for HD endoscopic images have been proposed, but these are computationally expensive, making them unfit for clinical use requiring real- time analysis. In this paper, we present a novel approach for early esophageal cancer detection using Transfer Learning with CNNs. Given the small amount of annotated data, CNN Codes are applied, where intermediate layers of the net- work are used as features for conventional classifiers. Various classifiers are combined with four of the most widely-used networks. Additionally, sliding windows are used to obtain a coarse-grained annotation indicating any possible cancerous regions. This approach outperforms the current state-of-the-art with a frame-based AUC of 0.92, while allowing both near real-time prediction and annotation at 2 fps, in a MATLAB-based framework.",
author = "{van Riel}, Sjors and {van der Sommen}, Fons and Sveta Zinger and Schoon, {Erik J.} and {de With}, {Peter H.N.}",
year = "2018",
month = "10",
doi = "10.1109/ICIP.2018.8451771",
language = "English",
isbn = "978-1-4799-7062-9",
pages = "1383--1387",
booktitle = "Proc. IEEE International Conference on Image Processing (ICIP)",
publisher = "IEEE Computer Society",
address = "United States",

}

van Riel, S, van der Sommen, F, Zinger, S, Schoon, EJ & de With, PHN 2018, Automatic detection of early esophageal cancer with CNNS using transfer learning. in Proc. IEEE International Conference on Image Processing (ICIP)., 8451771, IEEE Computer Society, Piscataway, pp. 1383-1387, 25th IEEE International Conference on Image Processing (ICIP 2018), Athens, Greece, 7/10/18. https://doi.org/10.1109/ICIP.2018.8451771

Automatic detection of early esophageal cancer with CNNS using transfer learning. / van Riel, Sjors; van der Sommen, Fons; Zinger, Sveta; Schoon, Erik J.; de With, Peter H.N.

Proc. IEEE International Conference on Image Processing (ICIP). Piscataway : IEEE Computer Society, 2018. p. 1383-1387 8451771.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Automatic detection of early esophageal cancer with CNNS using transfer learning

AU - van Riel, Sjors

AU - van der Sommen, Fons

AU - Zinger, Sveta

AU - Schoon, Erik J.

AU - de With, Peter H.N.

PY - 2018/10

Y1 - 2018/10

N2 - The incidence of Esophageal Adenocarcinoma (EAC), a form of esophageal cancer, has rapidly increased in recent years. Dysplastic tissue can be removed endoscopically at an early stage, and since survival chances of patients are limited at later stages of the disease, early detection is of key impor- tance. Recently, several CAD systems for HD endoscopic images have been proposed, but these are computationally expensive, making them unfit for clinical use requiring real- time analysis. In this paper, we present a novel approach for early esophageal cancer detection using Transfer Learning with CNNs. Given the small amount of annotated data, CNN Codes are applied, where intermediate layers of the net- work are used as features for conventional classifiers. Various classifiers are combined with four of the most widely-used networks. Additionally, sliding windows are used to obtain a coarse-grained annotation indicating any possible cancerous regions. This approach outperforms the current state-of-the-art with a frame-based AUC of 0.92, while allowing both near real-time prediction and annotation at 2 fps, in a MATLAB-based framework.

AB - The incidence of Esophageal Adenocarcinoma (EAC), a form of esophageal cancer, has rapidly increased in recent years. Dysplastic tissue can be removed endoscopically at an early stage, and since survival chances of patients are limited at later stages of the disease, early detection is of key impor- tance. Recently, several CAD systems for HD endoscopic images have been proposed, but these are computationally expensive, making them unfit for clinical use requiring real- time analysis. In this paper, we present a novel approach for early esophageal cancer detection using Transfer Learning with CNNs. Given the small amount of annotated data, CNN Codes are applied, where intermediate layers of the net- work are used as features for conventional classifiers. Various classifiers are combined with four of the most widely-used networks. Additionally, sliding windows are used to obtain a coarse-grained annotation indicating any possible cancerous regions. This approach outperforms the current state-of-the-art with a frame-based AUC of 0.92, while allowing both near real-time prediction and annotation at 2 fps, in a MATLAB-based framework.

U2 - 10.1109/ICIP.2018.8451771

DO - 10.1109/ICIP.2018.8451771

M3 - Conference contribution

SN - 978-1-4799-7062-9

SP - 1383

EP - 1387

BT - Proc. IEEE International Conference on Image Processing (ICIP)

PB - IEEE Computer Society

CY - Piscataway

ER -

van Riel S, van der Sommen F, Zinger S, Schoon EJ, de With PHN. Automatic detection of early esophageal cancer with CNNS using transfer learning. In Proc. IEEE International Conference on Image Processing (ICIP). Piscataway: IEEE Computer Society. 2018. p. 1383-1387. 8451771 https://doi.org/10.1109/ICIP.2018.8451771