Combining deep learning and hand-crafted features for skin lesion classification

Tomáš Majtner, Sule Yildirim-Yayilgan, Jon Yngve Hardeberg

Publikation: Bidrag til bog/antologi/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

Resumé

Melanoma is one of the most lethal forms of skin cancer. It occurs on the skin surface and develops from cells known as melanocytes. The same cells are also responsible for benign lesions commonly known as moles, which are visually similar to melanoma in its early stage. If melanoma is treated correctly, it is very often curable. Currently, much research is concentrated on the automated recognition of melanomas. In this paper, we propose an automated melanoma recognition system, which is based on deep learning method combined with so called hand-crafted RSurf features and Local Binary Patterns. The experimental evaluation on a large publicly available dataset demonstrates high classification accuracy, sensitivity, and specificity of our proposed approach when it is compared with other classifiers on the same dataset.

OriginalsprogEngelsk
Titel2016 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016
RedaktørerMatti Pietikainen, Abdenour Hadid, Miguel Bordallo Lopez
ForlagIEEE
Publikationsdato17. jan. 2017
Artikelnummer7821017
ISBN (Elektronisk)9781467389105
DOI
StatusUdgivet - 17. jan. 2017
Begivenhed6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016 - Oulu, Finland
Varighed: 12. dec. 201615. dec. 2016

Konference

Konference6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016
LandFinland
ByOulu
Periode12/12/201615/12/2016

Fingeraftryk

Skin
Classifiers
Deep learning

Citer dette

Majtner, T., Yildirim-Yayilgan, S., & Hardeberg, J. Y. (2017). Combining deep learning and hand-crafted features for skin lesion classification. I M. Pietikainen, A. Hadid, & M. B. Lopez (red.), 2016 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016 [7821017] IEEE. https://doi.org/10.1109/IPTA.2016.7821017
Majtner, Tomáš ; Yildirim-Yayilgan, Sule ; Hardeberg, Jon Yngve. / Combining deep learning and hand-crafted features for skin lesion classification. 2016 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016. red. / Matti Pietikainen ; Abdenour Hadid ; Miguel Bordallo Lopez. IEEE, 2017.
@inproceedings{51324f3d65904a61b034e5f3f1ab4964,
title = "Combining deep learning and hand-crafted features for skin lesion classification",
abstract = "Melanoma is one of the most lethal forms of skin cancer. It occurs on the skin surface and develops from cells known as melanocytes. The same cells are also responsible for benign lesions commonly known as moles, which are visually similar to melanoma in its early stage. If melanoma is treated correctly, it is very often curable. Currently, much research is concentrated on the automated recognition of melanomas. In this paper, we propose an automated melanoma recognition system, which is based on deep learning method combined with so called hand-crafted RSurf features and Local Binary Patterns. The experimental evaluation on a large publicly available dataset demonstrates high classification accuracy, sensitivity, and specificity of our proposed approach when it is compared with other classifiers on the same dataset.",
keywords = "Convolutional Neural Network, Local Binary Patterns, RSurf Features, Skin Lesion Classification, SVM",
author = "Tom{\'a}š Majtner and Sule Yildirim-Yayilgan and Hardeberg, {Jon Yngve}",
year = "2017",
month = "1",
day = "17",
doi = "10.1109/IPTA.2016.7821017",
language = "English",
editor = "Matti Pietikainen and Abdenour Hadid and Lopez, {Miguel Bordallo}",
booktitle = "2016 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016",
publisher = "IEEE",
address = "United States",

}

Majtner, T, Yildirim-Yayilgan, S & Hardeberg, JY 2017, Combining deep learning and hand-crafted features for skin lesion classification. i M Pietikainen, A Hadid & MB Lopez (red), 2016 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016., 7821017, IEEE, 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016, Oulu, Finland, 12/12/2016. https://doi.org/10.1109/IPTA.2016.7821017

Combining deep learning and hand-crafted features for skin lesion classification. / Majtner, Tomáš; Yildirim-Yayilgan, Sule; Hardeberg, Jon Yngve.

2016 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016. red. / Matti Pietikainen; Abdenour Hadid; Miguel Bordallo Lopez. IEEE, 2017. 7821017.

Publikation: Bidrag til bog/antologi/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

TY - GEN

T1 - Combining deep learning and hand-crafted features for skin lesion classification

AU - Majtner, Tomáš

AU - Yildirim-Yayilgan, Sule

AU - Hardeberg, Jon Yngve

PY - 2017/1/17

Y1 - 2017/1/17

N2 - Melanoma is one of the most lethal forms of skin cancer. It occurs on the skin surface and develops from cells known as melanocytes. The same cells are also responsible for benign lesions commonly known as moles, which are visually similar to melanoma in its early stage. If melanoma is treated correctly, it is very often curable. Currently, much research is concentrated on the automated recognition of melanomas. In this paper, we propose an automated melanoma recognition system, which is based on deep learning method combined with so called hand-crafted RSurf features and Local Binary Patterns. The experimental evaluation on a large publicly available dataset demonstrates high classification accuracy, sensitivity, and specificity of our proposed approach when it is compared with other classifiers on the same dataset.

AB - Melanoma is one of the most lethal forms of skin cancer. It occurs on the skin surface and develops from cells known as melanocytes. The same cells are also responsible for benign lesions commonly known as moles, which are visually similar to melanoma in its early stage. If melanoma is treated correctly, it is very often curable. Currently, much research is concentrated on the automated recognition of melanomas. In this paper, we propose an automated melanoma recognition system, which is based on deep learning method combined with so called hand-crafted RSurf features and Local Binary Patterns. The experimental evaluation on a large publicly available dataset demonstrates high classification accuracy, sensitivity, and specificity of our proposed approach when it is compared with other classifiers on the same dataset.

KW - Convolutional Neural Network

KW - Local Binary Patterns

KW - RSurf Features

KW - Skin Lesion Classification

KW - SVM

UR - http://www.scopus.com/inward/record.url?scp=85013141078&partnerID=8YFLogxK

U2 - 10.1109/IPTA.2016.7821017

DO - 10.1109/IPTA.2016.7821017

M3 - Article in proceedings

BT - 2016 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016

A2 - Pietikainen, Matti

A2 - Hadid, Abdenour

A2 - Lopez, Miguel Bordallo

PB - IEEE

ER -

Majtner T, Yildirim-Yayilgan S, Hardeberg JY. Combining deep learning and hand-crafted features for skin lesion classification. I Pietikainen M, Hadid A, Lopez MB, red., 2016 6th International Conference on Image Processing Theory, Tools and Applications, IPTA 2016. IEEE. 2017. 7821017 https://doi.org/10.1109/IPTA.2016.7821017