Melola Discovery

← Melola Discovery

Jurnal Tanpa peringkat SINTA

A Deep Learning Model Comparation for Diabetic Retinopathy Image Classification

Scientific Journal of Informatics · Vol. 12 No. 1 (2025) · hlm. 21–30

DOI
10.15294/sji.v12i1.20939
ISSN
2407-7658, 2460-0040
Terbit
2025-04-19
Penerbit
Universitas Negeri Semarang
Bahasa
eng

Abstrak

Purpose: This study compares the performance of various deep learning models for diabetic retinopathy (DR) classification, emphasizing the impact of different optimization functions. Early detection of DR is vital for preventing blindness, and the research investigates how optimization functions influence the classification accuracy and efficiency of several convolutional neural networks (CNNs). This study fills a gap in the existing literature by examining how optimization functions affect model performance in conjunction with architectural considerations. Methods: This paper uses the APTOS 2019 dataset, which comprises 3,663 retinal fundus images classified into five classes of diabetic retinopathy severity. Four CNN-based models, including CNN, ResNet50, DenseNet121, and EfficientNet B0, were trained using five optimization techniques: Adam, SGD, RMSProp, AdamW, and NAdam. The performance of the experimental scenarios was evaluated through accuracy, precision, recall, F1-score, training duration, and model size. Result: EfficientNet B0 demonstrated superior computational efficiency with a minimal model size of 16.16 MB. Subsequently, DenseNet121 with the SGD optimizer achieved the highest test accuracy of 96.86%. The experimental results indicate that the optimizer significantly influences model performance. AdamW and NAdam yield superior outcomes for deeper architectures such as ResNet50 and DenseNet121. Novelty: This paper offers an analytical examination of deep learning models and optimization techniques for DR classification, helping to clarify the trade-offs between computational efficiency and classification performance. The findings contribute to the development of more accurate and efficient DR detection systems, which could be utilized in real-world, resource-limited settings.

Kata kunci

Jaringan kolaborasi para penulis

14 penulis · 25 tautan kepenulisan bersama · 10 lintas organisasi

Arahkan kursor atau ketuk sebuah nama untuk menyorot rekan penulisnya. Klik dua kali untuk membuka halaman penulis.

  • Telkom University
  • Institut Teknologi Sepuluh Nopember
  • National Taiwan University of Science and Technology
  • University of Pembangunan Nasional Veteran Jawa Timur
  • Universitas Gadjah Mada
  • Afiliasi tidak diketahui
  • Organisasi berbeda
  • Organisasi sama
  • Afiliasi tidak diketahui
  • Garis lebih tebal = lebih banyak artikel bersama · lingkaran lebih besar = lebih banyak artikel
Tampilkan sebagai tabel
Rekan penulisOrganisasiArtikel bersama
Muhammad Ihsan DiputraInstitut Teknologi Sepuluh Nopember1
Ahmad Akbar MegantaraInstitut Teknologi Sepuluh Nopember1
Didik PurwantoNational Taiwan University of Science and Technology1
Aprilia Dewi ArdiyantiUniversity of Pembangunan Nasional Veteran Jawa Timur1
Allif Rosyidy HilmiAfiliasi tidak diketahui1
Yofinda Eka SetiawanAfiliasi tidak diketahui1
Aslam Chitami Priawan SiregarAfiliasi tidak diketahui1
Muhammad Dzulfikar FauziUniversitas Gadjah Mada1
Ilham YuliantoTelkom University1
Alqis RausanfitaInstitut Teknologi Sepuluh Nopember1
Muhammad Harits Shofwan AdaniTelkom University1

Afiliasi berasal dari OpenAlex; metadata OAI-PMH jurnal tidak memuatnya. Penulis tanpa ID OpenAlex dicocokkan berdasarkan nama, sehingga dua orang bisa tergabung atau satu orang bisa terpisah.

Sumber metadata: OAI-PMH (metadata jurnal) https://journal.unnes.ac.id/journals/sji/oai, pertama di-harvest 6 Nov 2026, terakhir terlihat 8 Nov 2026.
Metadata asli (oai_dc)
<oai_dc:dc xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/  http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
	<dc:title xml:lang="en">A Deep Learning Model Comparation for Diabetic Retinopathy Image Classification</dc:title>
	<dc:creator xml:lang="en">Mustaqim, Tanzilal </dc:creator>
	<dc:creator xml:lang="en">Safitri, Pima Hani</dc:creator>
	<dc:creator xml:lang="en">Muhajir, Daud</dc:creator>
	<dc:subject xml:lang="en">Diabetic retinopathy</dc:subject>
	<dc:subject xml:lang="en">Deep learning</dc:subject>
	<dc:subject xml:lang="en">Image classification</dc:subject>
	<dc:subject xml:lang="en">Convolutional neural networks (CNN)</dc:subject>
	<dc:subject xml:lang="en">Optimization algorithms</dc:subject>
	<dc:description xml:lang="en">Purpose: This study compares the performance of various deep learning models for diabetic retinopathy (DR) classification, emphasizing the impact of different optimization functions. Early detection of DR is vital for preventing blindness, and the research investigates how optimization functions influence the classification accuracy and efficiency of several convolutional neural networks (CNNs). This study fills a gap in the existing literature by examining how optimization functions affect model performance in conjunction with architectural considerations.
Methods: This paper uses the APTOS 2019 dataset, which comprises 3,663 retinal fundus images classified into five classes of diabetic retinopathy severity. Four CNN-based models, including CNN, ResNet50, DenseNet121, and EfficientNet B0, were trained using five optimization techniques: Adam, SGD, RMSProp, AdamW, and NAdam. The performance of the experimental scenarios was evaluated through accuracy, precision, recall, F1-score, training duration, and model size.
Result: EfficientNet B0 demonstrated superior computational efficiency with a minimal model size of 16.16 MB. Subsequently, DenseNet121 with the SGD optimizer achieved the highest test accuracy of 96.86%. The experimental results indicate that the optimizer significantly influences model performance. AdamW and NAdam yield superior outcomes for deeper architectures such as ResNet50 and DenseNet121.
Novelty: This paper offers an analytical examination of deep learning models and optimization techniques for DR classification, helping to clarify the trade-offs between computational efficiency and classification performance. The findings contribute to the development of more accurate and efficient DR detection systems, which could be utilized in real-world, resource-limited settings.</dc:description>
	<dc:publisher xml:lang="en">Universitas Negeri Semarang</dc:publisher>
	<dc:date>2025-04-19</dc:date>
	<dc:type>info:eu-repo/semantics/article</dc:type>
	<dc:type>info:eu-repo/semantics/publishedVersion</dc:type>
	<dc:type xml:lang="en">Peer-reviewed Article</dc:type>
	<dc:format>application/pdf</dc:format>
	<dc:identifier>https://journal.unnes.ac.id/journals/sji/article/view/20939</dc:identifier>
	<dc:identifier>10.15294/sji.v12i1.20939</dc:identifier>
	<dc:source xml:lang="en">Scientific Journal of Informatics; Vol. 12 No. 1: Regular Issue; 21-30</dc:source>
	<dc:source>2407-7658</dc:source>
	<dc:source>2460-0040</dc:source>
	<dc:source>10.15294/sji.v12i1</dc:source>
	<dc:language>eng</dc:language>
	<dc:relation>https://journal.unnes.ac.id/journals/sji/article/view/20939/3642</dc:relation>
	<dc:rights xml:lang="en">Copyright (c) 2025 Scientific Journal of Informatics</dc:rights>
</oai_dc:dc>