Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset

Maria  Tariq; Vasile Palade; YingLiang Ma

doi:10.1007/978-981-16-6775-6_8

Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset

Maria Tariq, Vasile Palade, YingLiang Ma

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

1 Citation (Scopus)

Abstract

Severe stages of diabetes can eventually lead to an eye condition called diabetic retinopathy. It is one of the leading causes of temporary visual disability and permanent blindness. There is no cure for this disease other than a proper treatment in the early stages. Five stages of diabetic retinopathy are discussed in this paper that need to be detected followed by a proper treatment. Transfer learning is used to detect the grades of diabetic retinopathy in eye fundus images, without training from scratch. The Kaggle EyePACS dataset is one of the largest datasets available publicly for experimentation. In our work, an extensive study on the Kaggle EyePACS dataset is carried out using the pre-trained models ResNet50 and DenseNet121. The Aptos dataset is also used in comparison with this dataset to examine the performance of the pre-trained models. Different experiments are performed to analyze the images from the different classes in the Kaggle EyePACS dataset. This dataset has significant challenges including image noise, imbalanced classes, and incorrect annotations. Our work highlights potential problems within the dataset and the conflicts between the classes. A clustering technique is used to get informative images from the normal class to improve the model’s accuracy to 70%.

Original language	English
Title of host publication	Medical Imaging and Computer-Aided Diagnosis
Subtitle of host publication	Proceedings of 2022 International Conference on Medical Imaging and Computer-Aided Diagnosis (MICAD 2022)
Editors	Ruidan Su, Yudong Zhang, Han Liu, Alejandro F. Frangi
Publisher	Springer
Pages	89–99
Number of pages	11
ISBN (Electronic)	978-981-16-6775-6
ISBN (Print)	978-981-16-6774-9
DOIs	https://doi.org/10.1007/978-981-16-6775-6_8
Publication status	Published - 20 Dec 2023
Event	The 3rd International Conference on Medical Imaging and Computer-Aided Diagnosis - University of Leicester, Leicester, United Kingdom Duration: 20 Nov 2022 → 21 Nov 2022 https://www.micad.org/

Publication series

Name	Lecture Notes in Electrical Engineering (LNEE)
Volume	810

Conference

Conference	The 3rd International Conference on Medical Imaging and Computer-Aided Diagnosis
Abbreviated title	MICAD2022
Country/Territory	United Kingdom
City	Leicester
Period	20/11/22 → 21/11/22
Internet address	https://www.micad.org/

Keywords

Deep learning
Fine tuning
Kaggle EyePACS
Pre-trained models
Transfer learning

UN SDGs

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

Access to Document

10.1007/978-981-16-6775-6_8

Cite this

Tariq, M., Palade, V., & Ma, Y. (2023). Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset. In R. Su, Y. Zhang, H. Liu, & A. F. Frangi (Eds.), Medical Imaging and Computer-Aided Diagnosis: Proceedings of 2022 International Conference on Medical Imaging and Computer-Aided Diagnosis (MICAD 2022) (pp. 89–99). (Lecture Notes in Electrical Engineering (LNEE); Vol. 810). Springer. https://doi.org/10.1007/978-981-16-6775-6_8

Tariq, Maria ; Palade, Vasile ; Ma, YingLiang. / Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset. Medical Imaging and Computer-Aided Diagnosis: Proceedings of 2022 International Conference on Medical Imaging and Computer-Aided Diagnosis (MICAD 2022). editor / Ruidan Su ; Yudong Zhang ; Han Liu ; Alejandro F. Frangi. Springer, 2023. pp. 89–99 (Lecture Notes in Electrical Engineering (LNEE)).

@inproceedings{c656281fb7ec4a67aea467e7c936b749,

title = "Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset",

abstract = "Severe stages of diabetes can eventually lead to an eye condition called diabetic retinopathy. It is one of the leading causes of temporary visual disability and permanent blindness. There is no cure for this disease other than a proper treatment in the early stages. Five stages of diabetic retinopathy are discussed in this paper that need to be detected followed by a proper treatment. Transfer learning is used to detect the grades of diabetic retinopathy in eye fundus images, without training from scratch. The Kaggle EyePACS dataset is one of the largest datasets available publicly for experimentation. In our work, an extensive study on the Kaggle EyePACS dataset is carried out using the pre-trained models ResNet50 and DenseNet121. The Aptos dataset is also used in comparison with this dataset to examine the performance of the pre-trained models. Different experiments are performed to analyze the images from the different classes in the Kaggle EyePACS dataset. This dataset has significant challenges including image noise, imbalanced classes, and incorrect annotations. Our work highlights potential problems within the dataset and the conflicts between the classes. A clustering technique is used to get informative images from the normal class to improve the model{\textquoteright}s accuracy to 70%.",

keywords = "Deep learning, Fine tuning, Kaggle EyePACS, Pre-trained models, Transfer learning",

author = "Maria Tariq and Vasile Palade and YingLiang Ma",

year = "2023",

month = dec,

day = "20",

doi = "10.1007/978-981-16-6775-6_8",

language = "English",

isbn = "978-981-16-6774-9",

series = "Lecture Notes in Electrical Engineering (LNEE)",

publisher = "Springer",

pages = "89–99",

editor = "Ruidan Su and Yudong Zhang and Han Liu and Frangi, {Alejandro F.}",

booktitle = "Medical Imaging and Computer-Aided Diagnosis",

address = "Germany",

note = "The 3rd International Conference on Medical Imaging and Computer-Aided Diagnosis, MICAD2022 ; Conference date: 20-11-2022 Through 21-11-2022",

url = "https://www.micad.org/",

}

Tariq, M, Palade, V & Ma, Y 2023, Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset. in R Su, Y Zhang, H Liu & AF Frangi (eds), Medical Imaging and Computer-Aided Diagnosis: Proceedings of 2022 International Conference on Medical Imaging and Computer-Aided Diagnosis (MICAD 2022). Lecture Notes in Electrical Engineering (LNEE), vol. 810, Springer, pp. 89–99, The 3rd International Conference on Medical Imaging and Computer-Aided Diagnosis, Leicester, United Kingdom, 20/11/22. https://doi.org/10.1007/978-981-16-6775-6_8

Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset. / Tariq, Maria ; Palade, Vasile; Ma, YingLiang.

Medical Imaging and Computer-Aided Diagnosis: Proceedings of 2022 International Conference on Medical Imaging and Computer-Aided Diagnosis (MICAD 2022). ed. / Ruidan Su; Yudong Zhang; Han Liu; Alejandro F. Frangi. Springer, 2023. p. 89–99 (Lecture Notes in Electrical Engineering (LNEE); Vol. 810).

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

TY - GEN

T1 - Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset

AU - Tariq, Maria

AU - Palade, Vasile

AU - Ma, YingLiang

PY - 2023/12/20

Y1 - 2023/12/20

N2 - Severe stages of diabetes can eventually lead to an eye condition called diabetic retinopathy. It is one of the leading causes of temporary visual disability and permanent blindness. There is no cure for this disease other than a proper treatment in the early stages. Five stages of diabetic retinopathy are discussed in this paper that need to be detected followed by a proper treatment. Transfer learning is used to detect the grades of diabetic retinopathy in eye fundus images, without training from scratch. The Kaggle EyePACS dataset is one of the largest datasets available publicly for experimentation. In our work, an extensive study on the Kaggle EyePACS dataset is carried out using the pre-trained models ResNet50 and DenseNet121. The Aptos dataset is also used in comparison with this dataset to examine the performance of the pre-trained models. Different experiments are performed to analyze the images from the different classes in the Kaggle EyePACS dataset. This dataset has significant challenges including image noise, imbalanced classes, and incorrect annotations. Our work highlights potential problems within the dataset and the conflicts between the classes. A clustering technique is used to get informative images from the normal class to improve the model’s accuracy to 70%.

AB - Severe stages of diabetes can eventually lead to an eye condition called diabetic retinopathy. It is one of the leading causes of temporary visual disability and permanent blindness. There is no cure for this disease other than a proper treatment in the early stages. Five stages of diabetic retinopathy are discussed in this paper that need to be detected followed by a proper treatment. Transfer learning is used to detect the grades of diabetic retinopathy in eye fundus images, without training from scratch. The Kaggle EyePACS dataset is one of the largest datasets available publicly for experimentation. In our work, an extensive study on the Kaggle EyePACS dataset is carried out using the pre-trained models ResNet50 and DenseNet121. The Aptos dataset is also used in comparison with this dataset to examine the performance of the pre-trained models. Different experiments are performed to analyze the images from the different classes in the Kaggle EyePACS dataset. This dataset has significant challenges including image noise, imbalanced classes, and incorrect annotations. Our work highlights potential problems within the dataset and the conflicts between the classes. A clustering technique is used to get informative images from the normal class to improve the model’s accuracy to 70%.

KW - Deep learning

KW - Fine tuning

KW - Kaggle EyePACS

KW - Pre-trained models

KW - Transfer learning

UR - http://www.scopus.com/inward/record.url?scp=85180811473&partnerID=8YFLogxK

U2 - 10.1007/978-981-16-6775-6_8

DO - 10.1007/978-981-16-6775-6_8

M3 - Conference contribution

SN - 978-981-16-6774-9

T3 - Lecture Notes in Electrical Engineering (LNEE)

SP - 89

EP - 99

BT - Medical Imaging and Computer-Aided Diagnosis

A2 - Su, Ruidan

A2 - Zhang, Yudong

A2 - Liu, Han

A2 - Frangi, Alejandro F.

PB - Springer

T2 - The 3rd International Conference on Medical Imaging and Computer-Aided Diagnosis

Y2 - 20 November 2022 through 21 November 2022

ER -

Tariq M, Palade V, Ma Y. Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset. In Su R, Zhang Y, Liu H, Frangi AF, editors, Medical Imaging and Computer-Aided Diagnosis: Proceedings of 2022 International Conference on Medical Imaging and Computer-Aided Diagnosis (MICAD 2022). Springer. 2023. p. 89–99. (Lecture Notes in Electrical Engineering (LNEE)). doi: 10.1007/978-981-16-6775-6_8

Transfer Learning based Classification of Diabetic Retinopathy on the Kaggle EyePACS dataset

Abstract

Publication series

Conference

Keywords

UN SDGs

Access to Document

Other files and links

Embargoed Document

Cite this