Diyabetik Retinopati Teşhisi için Fundus Görüntülerinin Derin Öğrenme Tabanlı Sınıflandırılması

Özçelik, Yusuf Bahri; Altan, Aytaç

doi:10.31590/ejosat.1011806

Cited by 10 publications

(1 citation statement)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, these screenings can be timeconsuming and require expertise. To address this challenge and support healthcare professionals, the utilization of rapid and reliable computer-aided automated screening systems is increasingly growing [5].…”

Section: Introductionmentioning

confidence: 99%

Ön Eğitimli Modeller ve Özellik Seçiminin Rolü: Diyabetik Retinopati Tanısında Yapay Zeka Tabanlı Yaklaşım

KAYA,

TASCİ

2023

Turkish Journal of Science and Technology

View full text Add to dashboard Cite

Diabetic retinopathy is a significant complication occurring in the retina of the eye as a result of prolonged diabetes. When not detected early, this condition can lead to vision loss. Advanced image processing techniques and artificial intelligence algorithms have enhanced the possibilities of early diagnosis and treatment. This article discusses current advancements in artificial intelligence-based diabetic retinopathy detection and explores future possibilities in this field. In the experimental studies of the article, the Kaggle Aptos 2019 dataset was utilized. This dataset comprises 5 classes and a total of 3662 images. The class distribution is as follows: No DR (No Diabetic Retinopathy): 1805, Mild: 370, Moderate: 999, Severe: 193, Proliferative DR: 295. The study consists of four fundamental stages. These stages are (1) Feature extraction from VGG16 and VGG19 pretrained models, (2) Feature selection using NCA, Relieff, and Chi2, (3) Classification with Support Vector Machine classifier, (4) Iterative Majority Voting. Using the proposed method, a high accuracy of 99.18% is achieved. Furthermore, sensitivity of 100% for the No DR class, sensitivity of 100% for the Moderate class, sensitivity of 98.80% for the Severe class, and an F1-Score of 99.89% for the No DR class are obtained. This study demonstrates the effective utilization of machine learning methods in diabetic retinopathy diagnosis. The experimental results underscore the significant contributions of diabetic retinopathy patients' diagnosis and treatment processes.

show abstract

Section: Introductionmentioning

confidence: 99%

Ön Eğitimli Modeller ve Özellik Seçiminin Rolü: Diyabetik Retinopati Tanısında Yapay Zeka Tabanlı Yaklaşım

KAYA,

TASCİ

2023

Turkish Journal of Science and Technology

View full text Add to dashboard Cite

show abstract

DIAGNOSIS AND CLASSIFICATION OF DIABETIC RETINOPATHY WITH YOLOv8-BASED DEEP LEARNING MODEL

Şanver,

Saygılı

2024

Kahramanmaraş Sütçü İmam Üniversitesi Mühendislik Bilimleri Dergisi

View full text Add to dashboard Cite

Diabetic retinopathy is a complication of diabetes that affects the eyes. High blood sugar levels damage the vessels of the retina, damaging the light-sensing cells in the eye, and can cause vision loss and, in severe cases, blindness. Deep learning models are powerful tools that can process and learn large data sets, and if used in the diagnosis of diabetes and diabetic retinopathy, they can benefit the early diagnosis of the disease. Deep learning enables early and high-accuracy detection of diabetic retinopathy symptoms with high sensitivity and specificity, as well as minimizing errors made by experts. In this study, we aimed to detect and classify diabetic retinopathy using the YOLOv8 (You Only Look Once) model, one of the CNN (convolutional neural network) architectures. The experimental studies were conducted with two different CPUs and two different GPUs. As a result of the experimental studies, the highest accuracy value was obtained as 84.91% with GPU1, and the average accuracy across the four different methods (CPU1, CPU2, GPU1, GPU2) was 83.82%.

show abstract

Görüntü Fi̇ltreleme İle Deneti̇msi̇z Reti̇na Damar Bölütlenmesi̇ İçi̇n Parametre Eni̇yi̇leşti̇ri̇lmesi̇

Yakut

Ulukaya

2022

Mühendislik Bilimleri Ve Tasarım Dergisi

View full text Add to dashboard Cite

For the detection and evaluation of eye disorders, retinal pictures are obtained in a digital environment with a customized camera system called the fundus. Due to various noises and unsharp contrast, it is difficult to detect the vessels in the eye by specialists, and this can make it difficult for specialists to diagnose. In this study, unsupervised image processing-based mathematical morphology and Coye filtering, and connected component analysis approaches were used to increase the success of retinal vascular segmentation from fundus images. In addition, retinal images are preprocessed for noise reduction and increased contrast. Parameter optimization was performed to increase the success of unsupervised image processing-based approaches. In the contrast limited adaptive histogram equalization (CLAHE) method, which is frequently used in image processing, the most appropriate upper limit value for contrast on color retinal images was investigated. The presented approach tested on the DRIVE and STARE datasets available to researchers. Compared to previous unsupervised learning studies, some metrics were at par with the literature and some metrics were more successful.

show abstract

Diyabetik Retinopati Teşhisi için Fundus Görüntülerinin Derin Öğrenme Tabanlı Sınıflandırılması

Cited by 10 publications

References 30 publications

Ön Eğitimli Modeller ve Özellik Seçiminin Rolü: Diyabetik Retinopati Tanısında Yapay Zeka Tabanlı Yaklaşım

Ön Eğitimli Modeller ve Özellik Seçiminin Rolü: Diyabetik Retinopati Tanısında Yapay Zeka Tabanlı Yaklaşım

DIAGNOSIS AND CLASSIFICATION OF DIABETIC RETINOPATHY WITH YOLOv8-BASED DEEP LEARNING MODEL

Görüntü Fi̇ltreleme İle Deneti̇msi̇z Reti̇na Damar Bölütlenmesi̇ İçi̇n Parametre Eni̇yi̇leşti̇ri̇lmesi̇

Contact Info

Product

Resources

About