Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification

Mayar Lotfy; Raed M. Shubair; Nassir Navab; Shadi Albarqouni

doi:10.1109/ICECTA48151.2019.8959770

Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification

Mayar Lotfy, Raed M. Shubair, Nassir Navab, Shadi Albarqouni

Informatics 16 - Chair of Computer Aided Medical Procedures

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

11 Scopus citations

Abstract

This paper develops an approach based on deep learning for the classifications of a common critical type of bone fractures, namely proximal femur. The performance of the state-of-the-art deep learning architecture, DenseNet, is investigated along with a recently introduced loss function, focal loss, to address the problem of imbalanced classes. Quantitative assessment is carried out on a real dataset consisting of 1347 X-ray images. Results demonstrate that the proposed deep learning approach utilizing focal loss show better performance for the fracture detection case and comparable results for the classification scenarios.

Original language	English
Title of host publication	2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781728155326
DOIs	https://doi.org/10.1109/ICECTA48151.2019.8959770
State	Published - Nov 2019
Event	2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019 - Ras Al Khaimah, United Arab Emirates Duration: 19 Nov 2019 → 21 Nov 2019

Publication series

Name	2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019

Conference

Conference	2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019
Country/Territory	United Arab Emirates
City	Ras Al Khaimah
Period	19/11/19 → 21/11/19

Keywords

Deep Learning
DenseNet
Focal loss
Proximal femur fractures

Access to Document

10.1109/ICECTA48151.2019.8959770

Cite this

Lotfy, M., Shubair, R. M., Navab, N., & Albarqouni, S. (2019). Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification. In 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019 Article 8959770 (2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICECTA48151.2019.8959770

Lotfy, Mayar ; Shubair, Raed M. ; Navab, Nassir et al. / Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification. 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019. Institute of Electrical and Electronics Engineers Inc., 2019. (2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019).

@inproceedings{fcdfe6f50bf14fdba588041e407cf32c,

title = "Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification",

abstract = "This paper develops an approach based on deep learning for the classifications of a common critical type of bone fractures, namely proximal femur. The performance of the state-of-the-art deep learning architecture, DenseNet, is investigated along with a recently introduced loss function, focal loss, to address the problem of imbalanced classes. Quantitative assessment is carried out on a real dataset consisting of 1347 X-ray images. Results demonstrate that the proposed deep learning approach utilizing focal loss show better performance for the fracture detection case and comparable results for the classification scenarios.",

keywords = "Deep Learning, DenseNet, Focal loss, Proximal femur fractures",

author = "Mayar Lotfy and Shubair, {Raed M.} and Nassir Navab and Shadi Albarqouni",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019 ; Conference date: 19-11-2019 Through 21-11-2019",

year = "2019",

month = nov,

doi = "10.1109/ICECTA48151.2019.8959770",

language = "English",

series = "2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019",

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

booktitle = "2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019",

}

Lotfy, M, Shubair, RM, Navab, N & Albarqouni, S 2019, Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification. in 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019., 8959770, 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019, Institute of Electrical and Electronics Engineers Inc., 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019, Ras Al Khaimah, United Arab Emirates, 19/11/19. https://doi.org/10.1109/ICECTA48151.2019.8959770

Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification. / Lotfy, Mayar; Shubair, Raed M.; Navab, Nassir et al.
2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019. Institute of Electrical and Electronics Engineers Inc., 2019. 8959770 (2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019).

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

TY - GEN

T1 - Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification

AU - Lotfy, Mayar

AU - Shubair, Raed M.

AU - Navab, Nassir

AU - Albarqouni, Shadi

PY - 2019/11

Y1 - 2019/11

N2 - This paper develops an approach based on deep learning for the classifications of a common critical type of bone fractures, namely proximal femur. The performance of the state-of-the-art deep learning architecture, DenseNet, is investigated along with a recently introduced loss function, focal loss, to address the problem of imbalanced classes. Quantitative assessment is carried out on a real dataset consisting of 1347 X-ray images. Results demonstrate that the proposed deep learning approach utilizing focal loss show better performance for the fracture detection case and comparable results for the classification scenarios.

AB - This paper develops an approach based on deep learning for the classifications of a common critical type of bone fractures, namely proximal femur. The performance of the state-of-the-art deep learning architecture, DenseNet, is investigated along with a recently introduced loss function, focal loss, to address the problem of imbalanced classes. Quantitative assessment is carried out on a real dataset consisting of 1347 X-ray images. Results demonstrate that the proposed deep learning approach utilizing focal loss show better performance for the fracture detection case and comparable results for the classification scenarios.

KW - Deep Learning

KW - DenseNet

KW - Focal loss

KW - Proximal femur fractures

UR - http://www.scopus.com/inward/record.url?scp=85078878505&partnerID=8YFLogxK

U2 - 10.1109/ICECTA48151.2019.8959770

DO - 10.1109/ICECTA48151.2019.8959770

M3 - Conference contribution

AN - SCOPUS:85078878505

T3 - 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019

BT - 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019

Y2 - 19 November 2019 through 21 November 2019

ER -

Lotfy M, Shubair RM, Navab N, Albarqouni S. Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification. In 2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019. Institute of Electrical and Electronics Engineers Inc. 2019. 8959770. (2019 International Conference on Electrical and Computing Technologies and Applications, ICECTA 2019). doi: 10.1109/ICECTA48151.2019.8959770

Investigation of Focal Loss in Deep Learning Models for Femur Fractures Classification

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this