D’OraCa: deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer

Jian Han Lim; Chun Shui Tan; Chee Seng Chan; Roshan Alex Welikala; Paolo Remagnino; Senthilmani Rajendran; Thomas George Kallarakkal; Rosnah Binti Zain; Ruwan Duminda Jayasinghe; Jyotsna Rimal; Alexander Ross Kerr; Rahmi Amtha; Karthikeya Patil; Wanninayake Mudiyanselage Tilakaratne; John Gibson; Sok Ching Cheong; Sarah Ann Barman

doi:10.1007/978-3-030-80432-9_31

D’OraCa: deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer

Jian Han Lim
, Chun Shui Tan
, Chee Seng Chan
, Roshan Alex Welikala
, Paolo Remagnino
, Senthilmani Rajendran
, Thomas George Kallarakkal
, Rosnah Binti Zain
, Ruwan Duminda Jayasinghe
, Jyotsna Rimal
, Alexander Ross Kerr
, Rahmi Amtha
, Karthikeya Patil
, Wanninayake Mudiyanselage Tilakaratne
, John Gibson
, Sok Ching Cheong
, Sarah Ann Barman

University of Malaya
Cancer Research Malaysia
MAHSA University
University of Peradeniya
B P Koirala Institute of Health Sciences
New York University
Trisakti University
JSS Academy of Higher Education & Research
University of Aberdeen

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

16 Downloads (Pure)

Abstract

Oral cancer is a major health issue among low- and middle-income countries due to the late diagnosis. Automated algorithms and tools have the potential to identify oral lesions for early detection of oral cancer. In this paper, we aim to develop a novel deep learning framework named D’OraCa to classify oral lesions using photographic images. We are the first to develop a mouth landmark detection model for the oral images and incorporate it into the oral lesion classification model as a guidance to improve the classification accuracy. We evaluate the performance of five different deep convolutional neural networks and MobileNetV2 was chosen as the feature extractor for our proposed mouth landmark detection model. Quantitative and qualitative results demonstrate the effectiveness of the mouth landmark detection model in guiding the classification model to classify the oral lesions into four different referral decision classes. We train our proposed mouth landmark model on a combination of five datasets, containing 221,565 images. Then, we train and evaluate our proposed classification model with mouth landmark guidance using 2,455 oral images. The results are consistent with clinicians and the
score of the classification model is improved to 61.68%.

Original language	English
Title of host publication	Medical Image Understanding and Analysis
Subtitle of host publication	25th Annual Conference, MIUA 2021, Oxford, United Kingdom, July 12–14, 2021, proceedings
Editors	Bartłomiej W. Papież, Mohammad Yaqub, Jianbo Jiao, Ana I. L. Namburete, J. Alison Noble
Place of Publication	Cham, Switzerland
Publisher	Springer Nature
Pages	408–422
Number of pages	13
ISBN (Electronic)	9783030804329
ISBN (Print)	9783030804312
DOIs	https://doi.org/10.1007/978-3-030-80432-9_31
Publication status	Published - 6 Jul 2021
Event	25th UK Conference on Medical Image Understanding and Analysis (MIUA2021) - Oxford, U.K. (held online) Duration: 12 Jul 2021 → 14 Jul 2021

Publication series

Name	Lecture Notes in Computer Science
Volume	12722
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	25th UK Conference on Medical Image Understanding and Analysis (MIUA2021)
Period	12/07/21 → 14/07/21

Bibliographical note

Organising Body: MIUA

Keywords

Cancer studies

Access to Document

10.1007/978-3-030-80432-9_31

Barman-S-49763-AAMAccepted author manuscript, 4.95 MBLicence: Unspecified

Cite this

Lim, J. H., Tan, C. S., Chan, C. S., Welikala, R. A., Remagnino, P., Rajendran, S., Kallarakkal, T. G., Zain, R. B., Jayasinghe, R. D., Rimal, J., Kerr, A. R., Amtha, R., Patil, K., Tilakaratne, W. M., Gibson, J., Cheong, S. C., & Barman, S. A. (2021). D’OraCa: deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer. In B. W. Papież, M. Yaqub, J. Jiao, A. I. L. Namburete, & J. A. Noble (Eds.), Medical Image Understanding and Analysis: 25th Annual Conference, MIUA 2021, Oxford, United Kingdom, July 12–14, 2021, proceedings (pp. 408–422). (Lecture Notes in Computer Science; Vol. 12722). Springer Nature. https://doi.org/10.1007/978-3-030-80432-9_31

Lim, Jian Han ; Tan, Chun Shui ; Chan, Chee Seng et al. / D’OraCa : deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer. Medical Image Understanding and Analysis: 25th Annual Conference, MIUA 2021, Oxford, United Kingdom, July 12–14, 2021, proceedings. editor / Bartłomiej W. Papież ; Mohammad Yaqub ; Jianbo Jiao ; Ana I. L. Namburete ; J. Alison Noble. Cham, Switzerland : Springer Nature, 2021. pp. 408–422 (Lecture Notes in Computer Science).

@inproceedings{18727c1a9d584a4aaf6c1f0e7e10a75d,

title = "D{\textquoteright}OraCa: deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer",

abstract = "Oral cancer is a major health issue among low- and middle-income countries due to the late diagnosis. Automated algorithms and tools have the potential to identify oral lesions for early detection of oral cancer. In this paper, we aim to develop a novel deep learning framework named D{\textquoteright}OraCa to classify oral lesions using photographic images. We are the first to develop a mouth landmark detection model for the oral images and incorporate it into the oral lesion classification model as a guidance to improve the classification accuracy. We evaluate the performance of five different deep convolutional neural networks and MobileNetV2 was chosen as the feature extractor for our proposed mouth landmark detection model. Quantitative and qualitative results demonstrate the effectiveness of the mouth landmark detection model in guiding the classification model to classify the oral lesions into four different referral decision classes. We train our proposed mouth landmark model on a combination of five datasets, containing 221,565 images. Then, we train and evaluate our proposed classification model with mouth landmark guidance using 2,455 oral images. The results are consistent with clinicians and the score of the classification model is improved to 61.68\%.",

keywords = "Cancer studies",

author = "Lim, \{Jian Han\} and Tan, \{Chun Shui\} and Chan, \{Chee Seng\} and Welikala, \{Roshan Alex\} and Paolo Remagnino and Senthilmani Rajendran and Kallarakkal, \{Thomas George\} and Zain, \{Rosnah Binti\} and Jayasinghe, \{Ruwan Duminda\} and Jyotsna Rimal and Kerr, \{Alexander Ross\} and Rahmi Amtha and Karthikeya Patil and Tilakaratne, \{Wanninayake Mudiyanselage\} and John Gibson and Cheong, \{Sok Ching\} and Barman, \{Sarah Ann\}",

note = "Organising Body: MIUA; 25th UK Conference on Medical Image Understanding and Analysis (MIUA2021) ; Conference date: 12-07-2021 Through 14-07-2021",

year = "2021",

month = jul,

day = "6",

doi = "10.1007/978-3-030-80432-9\_31",

language = "English",

isbn = "9783030804312",

series = "Lecture Notes in Computer Science",

publisher = "Springer Nature",

pages = "408{\textendash}422",

editor = "Papie{\.z}, \{Bart{\l}omiej W.\} and Mohammad Yaqub and Jianbo Jiao and Namburete, \{Ana I. L.\} and Noble, \{ J. Alison\}",

booktitle = "Medical Image Understanding and Analysis",

address = "Switzerland",

}

Lim, JH, Tan, CS, Chan, CS, Welikala, RA, Remagnino, P, Rajendran, S, Kallarakkal, TG, Zain, RB, Jayasinghe, RD, Rimal, J, Kerr, AR, Amtha, R, Patil, K, Tilakaratne, WM, Gibson, J, Cheong, SC & Barman, SA 2021, D’OraCa: deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer. in BW Papież, M Yaqub, J Jiao, AIL Namburete & JA Noble (eds), Medical Image Understanding and Analysis: 25th Annual Conference, MIUA 2021, Oxford, United Kingdom, July 12–14, 2021, proceedings. Lecture Notes in Computer Science, vol. 12722, Springer Nature, Cham, Switzerland, pp. 408–422, 25th UK Conference on Medical Image Understanding and Analysis (MIUA2021), 12/07/21. https://doi.org/10.1007/978-3-030-80432-9_31

D’OraCa: deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer. / Lim, Jian Han; Tan, Chun Shui; Chan, Chee Seng et al.
Medical Image Understanding and Analysis: 25th Annual Conference, MIUA 2021, Oxford, United Kingdom, July 12–14, 2021, proceedings. ed. / Bartłomiej W. Papież; Mohammad Yaqub; Jianbo Jiao; Ana I. L. Namburete; J. Alison Noble. Cham, Switzerland: Springer Nature, 2021. p. 408–422 (Lecture Notes in Computer Science; Vol. 12722).

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

TY - GEN

T1 - D’OraCa

T2 - 25th UK Conference on Medical Image Understanding and Analysis (MIUA2021)

AU - Lim, Jian Han

AU - Tan, Chun Shui

AU - Chan, Chee Seng

AU - Welikala, Roshan Alex

AU - Remagnino, Paolo

AU - Rajendran, Senthilmani

AU - Kallarakkal, Thomas George

AU - Zain, Rosnah Binti

AU - Jayasinghe, Ruwan Duminda

AU - Rimal, Jyotsna

AU - Kerr, Alexander Ross

AU - Amtha, Rahmi

AU - Patil, Karthikeya

AU - Tilakaratne, Wanninayake Mudiyanselage

AU - Gibson, John

AU - Cheong, Sok Ching

AU - Barman, Sarah Ann

N1 - Organising Body: MIUA

PY - 2021/7/6

Y1 - 2021/7/6

N2 - Oral cancer is a major health issue among low- and middle-income countries due to the late diagnosis. Automated algorithms and tools have the potential to identify oral lesions for early detection of oral cancer. In this paper, we aim to develop a novel deep learning framework named D’OraCa to classify oral lesions using photographic images. We are the first to develop a mouth landmark detection model for the oral images and incorporate it into the oral lesion classification model as a guidance to improve the classification accuracy. We evaluate the performance of five different deep convolutional neural networks and MobileNetV2 was chosen as the feature extractor for our proposed mouth landmark detection model. Quantitative and qualitative results demonstrate the effectiveness of the mouth landmark detection model in guiding the classification model to classify the oral lesions into four different referral decision classes. We train our proposed mouth landmark model on a combination of five datasets, containing 221,565 images. Then, we train and evaluate our proposed classification model with mouth landmark guidance using 2,455 oral images. The results are consistent with clinicians and the score of the classification model is improved to 61.68%.

AB - Oral cancer is a major health issue among low- and middle-income countries due to the late diagnosis. Automated algorithms and tools have the potential to identify oral lesions for early detection of oral cancer. In this paper, we aim to develop a novel deep learning framework named D’OraCa to classify oral lesions using photographic images. We are the first to develop a mouth landmark detection model for the oral images and incorporate it into the oral lesion classification model as a guidance to improve the classification accuracy. We evaluate the performance of five different deep convolutional neural networks and MobileNetV2 was chosen as the feature extractor for our proposed mouth landmark detection model. Quantitative and qualitative results demonstrate the effectiveness of the mouth landmark detection model in guiding the classification model to classify the oral lesions into four different referral decision classes. We train our proposed mouth landmark model on a combination of five datasets, containing 221,565 images. Then, we train and evaluate our proposed classification model with mouth landmark guidance using 2,455 oral images. The results are consistent with clinicians and the score of the classification model is improved to 61.68%.

KW - Cancer studies

U2 - 10.1007/978-3-030-80432-9_31

DO - 10.1007/978-3-030-80432-9_31

M3 - Conference contribution

SN - 9783030804312

T3 - Lecture Notes in Computer Science

SP - 408

EP - 422

BT - Medical Image Understanding and Analysis

A2 - Papież, Bartłomiej W.

A2 - Yaqub, Mohammad

A2 - Jiao, Jianbo

A2 - Namburete, Ana I. L.

A2 - Noble, J. Alison

PB - Springer Nature

CY - Cham, Switzerland

Y2 - 12 July 2021 through 14 July 2021

ER -

Lim JH, Tan CS, Chan CS, Welikala RA, Remagnino P, Rajendran S et al. D’OraCa: deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer. In Papież BW, Yaqub M, Jiao J, Namburete AIL, Noble JA, editors, Medical Image Understanding and Analysis: 25th Annual Conference, MIUA 2021, Oxford, United Kingdom, July 12–14, 2021, proceedings. Cham, Switzerland: Springer Nature. 2021. p. 408–422. (Lecture Notes in Computer Science). doi: 10.1007/978-3-030-80432-9_31

D’OraCa: deep learning-based classification of oral lesions with mouth landmark guidance for early detection of oral cancer

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this