A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification

Mohammed A. Al-Masni; Mugahed A. Al-Antari; Hye Min Park; Na Hyeon Park; Tae Seong Kim

doi:10.1109/ECBIOS.2019.8807441

A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification

Mohammed A. Al-Masni
, Mugahed A. Al-Antari
, Hye Min Park
, Na Hyeon Park
, Tae Seong Kim^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

25 Scopus citations

Abstract

Automated diagnosis of various skin lesion diseases through medical dermoscopy images is still a very challenging task. In this work, an integrated model for segmentation of skin lesion boundaries and classification of skin lesions is proposed by cascading novel deep learning networks. In the first stage, a novel full resolution convolutional networks (FrCN) is utilized to segment the boundaries of skin lesions from dermoscopy images. Then, the segmented lesions are passed into a deep residual networks (i.e., ResNet-50) for classification. The pre-segmentation process enables ResNet-50 to extract more specific and representative features from skin lesions and use them for improved classification. We have tested and evaluated our diagnostic deep model for skin lesions using the publicly available International Skin Imaging Collaboration (ISIC) 2017 challenge dataset which contains three different skin diseases: Benign, seborrheic keratosis, and melanoma. The integrated model exhibits its capability to segment the skin lesions with an overall accuracy of 94.03% and an average Jaccard similarity index of 77.11% via FrCN. Meanwhile, the overall prediction accuracy and F1-score of multiple skin lesions classification task via ResNet-50 achieved 81.57% and 75.75%, respectively. The integrated model could be utilized as a computer-aided diagnosis (CAD) system for dermatology.

Original language	English
Title of host publication	Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019
Editors	Teen-Hang Meen
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	95-98
Number of pages	4
ISBN (Electronic)	9781728107639
DOIs	https://doi.org/10.1109/ECBIOS.2019.8807441
State	Published - May 2019
Externally published	Yes
Event	2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019 - Okinawa, Japan Duration: 31 May 2019 → 3 Jun 2019

Publication series

Name	Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019

Conference

Conference	2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019
Country/Territory	Japan
City	Okinawa
Period	31/05/19 → 3/06/19

Bibliographical note

Publisher Copyright:
© 2019 IEEE.

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

Computer-Aided Diagnosis
Full Resolution Convolutional Network (FrCN)
Melanoma
Residual Networks
Segmentation and Classification of Skin Lesions

ASJC Scopus subject areas

Biomedical Engineering
Safety, Risk, Reliability and Quality
Building and Construction
Health Policy
Health(social science)
Strategy and Management
Renewable Energy, Sustainability and the Environment

Access to Document

10.1109/ECBIOS.2019.8807441

Cite this

Al-Masni, M. A., Al-Antari, M. A., Park, H. M., Park, N. H., & Kim, T. S. (2019). A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification. In T.-H. Meen (Ed.), Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019 (pp. 95-98). Article 8807441 (Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ECBIOS.2019.8807441

Al-Masni, Mohammed A. ; Al-Antari, Mugahed A. ; Park, Hye Min et al. / A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification. Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019. editor / Teen-Hang Meen. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 95-98 (Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019).

@inproceedings{b37db3663d214d84aae8765d3ca6e622,

title = "A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification",

abstract = "Automated diagnosis of various skin lesion diseases through medical dermoscopy images is still a very challenging task. In this work, an integrated model for segmentation of skin lesion boundaries and classification of skin lesions is proposed by cascading novel deep learning networks. In the first stage, a novel full resolution convolutional networks (FrCN) is utilized to segment the boundaries of skin lesions from dermoscopy images. Then, the segmented lesions are passed into a deep residual networks (i.e., ResNet-50) for classification. The pre-segmentation process enables ResNet-50 to extract more specific and representative features from skin lesions and use them for improved classification. We have tested and evaluated our diagnostic deep model for skin lesions using the publicly available International Skin Imaging Collaboration (ISIC) 2017 challenge dataset which contains three different skin diseases: Benign, seborrheic keratosis, and melanoma. The integrated model exhibits its capability to segment the skin lesions with an overall accuracy of 94.03\% and an average Jaccard similarity index of 77.11\% via FrCN. Meanwhile, the overall prediction accuracy and F1-score of multiple skin lesions classification task via ResNet-50 achieved 81.57\% and 75.75\%, respectively. The integrated model could be utilized as a computer-aided diagnosis (CAD) system for dermatology.",

keywords = "Computer-Aided Diagnosis, Full Resolution Convolutional Network (FrCN), Melanoma, Residual Networks, Segmentation and Classification of Skin Lesions",

author = "Al-Masni, \{Mohammed A.\} and Al-Antari, \{Mugahed A.\} and Park, \{Hye Min\} and Park, \{Na Hyeon\} and Kim, \{Tae Seong\}",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019 ; Conference date: 31-05-2019 Through 03-06-2019",

year = "2019",

month = may,

doi = "10.1109/ECBIOS.2019.8807441",

language = "English",

series = "Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "95--98",

editor = "Teen-Hang Meen",

booktitle = "Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019",

address = "United States",

}

Al-Masni, MA, Al-Antari, MA, Park, HM, Park, NH & Kim, TS 2019, A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification. in T-H Meen (ed.), Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019., 8807441, Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019, Institute of Electrical and Electronics Engineers Inc., pp. 95-98, 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019, Okinawa, Japan, 31/05/19. https://doi.org/10.1109/ECBIOS.2019.8807441

A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification. / Al-Masni, Mohammed A.; Al-Antari, Mugahed A.; Park, Hye Min et al.
Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019. ed. / Teen-Hang Meen. Institute of Electrical and Electronics Engineers Inc., 2019. p. 95-98 8807441 (Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification

AU - Al-Masni, Mohammed A.

AU - Al-Antari, Mugahed A.

AU - Park, Hye Min

AU - Park, Na Hyeon

AU - Kim, Tae Seong

PY - 2019/5

Y1 - 2019/5

N2 - Automated diagnosis of various skin lesion diseases through medical dermoscopy images is still a very challenging task. In this work, an integrated model for segmentation of skin lesion boundaries and classification of skin lesions is proposed by cascading novel deep learning networks. In the first stage, a novel full resolution convolutional networks (FrCN) is utilized to segment the boundaries of skin lesions from dermoscopy images. Then, the segmented lesions are passed into a deep residual networks (i.e., ResNet-50) for classification. The pre-segmentation process enables ResNet-50 to extract more specific and representative features from skin lesions and use them for improved classification. We have tested and evaluated our diagnostic deep model for skin lesions using the publicly available International Skin Imaging Collaboration (ISIC) 2017 challenge dataset which contains three different skin diseases: Benign, seborrheic keratosis, and melanoma. The integrated model exhibits its capability to segment the skin lesions with an overall accuracy of 94.03% and an average Jaccard similarity index of 77.11% via FrCN. Meanwhile, the overall prediction accuracy and F1-score of multiple skin lesions classification task via ResNet-50 achieved 81.57% and 75.75%, respectively. The integrated model could be utilized as a computer-aided diagnosis (CAD) system for dermatology.

AB - Automated diagnosis of various skin lesion diseases through medical dermoscopy images is still a very challenging task. In this work, an integrated model for segmentation of skin lesion boundaries and classification of skin lesions is proposed by cascading novel deep learning networks. In the first stage, a novel full resolution convolutional networks (FrCN) is utilized to segment the boundaries of skin lesions from dermoscopy images. Then, the segmented lesions are passed into a deep residual networks (i.e., ResNet-50) for classification. The pre-segmentation process enables ResNet-50 to extract more specific and representative features from skin lesions and use them for improved classification. We have tested and evaluated our diagnostic deep model for skin lesions using the publicly available International Skin Imaging Collaboration (ISIC) 2017 challenge dataset which contains three different skin diseases: Benign, seborrheic keratosis, and melanoma. The integrated model exhibits its capability to segment the skin lesions with an overall accuracy of 94.03% and an average Jaccard similarity index of 77.11% via FrCN. Meanwhile, the overall prediction accuracy and F1-score of multiple skin lesions classification task via ResNet-50 achieved 81.57% and 75.75%, respectively. The integrated model could be utilized as a computer-aided diagnosis (CAD) system for dermatology.

KW - Computer-Aided Diagnosis

KW - Full Resolution Convolutional Network (FrCN)

KW - Melanoma

KW - Residual Networks

KW - Segmentation and Classification of Skin Lesions

UR - https://www.scopus.com/pages/publications/85072025836

U2 - 10.1109/ECBIOS.2019.8807441

DO - 10.1109/ECBIOS.2019.8807441

M3 - Conference contribution

AN - SCOPUS:85072025836

T3 - Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019

SP - 95

EP - 98

BT - Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019

A2 - Meen, Teen-Hang

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019

Y2 - 31 May 2019 through 3 June 2019

ER -

Al-Masni MA, Al-Antari MA, Park HM, Park NH, Kim TS. A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification. In Meen TH, editor, Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 95-98. 8807441. (Proceedings of 2019 IEEE Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability, ECBIOS 2019). doi: 10.1109/ECBIOS.2019.8807441

A Deep Learning Model Integrating FrCN and Residual Convolutional Networks for Skin Lesion Segmentation and Classification

Abstract

Publication series

Conference

Bibliographical note

UN SDGs

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this