Classification of Diabetic Retinopathy Disease Levels by Extracting Topological Features Using Graph Neural Networks
Fuente:
Zenodo
Gespeichert in:
| 1. Verfasser: | |
|---|---|
| Format: | Recurso digital |
| Sprache: | Englisch |
| Veröffentlicht: |
Zenodo
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866902123412717568 |
|---|---|
| author | UtilitasMathematica |
| author_facet | UtilitasMathematica |
| contents | <p>Diabetic retinopathy (DR) is one of the most common causes of blindness in the world”. It has to be diagnosed quickly and accurately so that treatment may begin as soon as possible. Manual examination of fundus photographs by doctors is prone to mistakes and is labor-intensive. “Using computer-assisted methods, especially Convolutional Neural Networks (CNNs), to automate DR diagnosis seems promising”. The goal of this research is to improve the processing of retinal pictures by using a “Graph Convolutional Neural network (GCNN)” to classify the severity of diseases. GCNNs increase feature extraction by using topological correlations in pictures. This makes classification results more accurate. The suggested GCNN model works well, as shown by evaluation measures including “accuracy, precision, recall, and F1-score”. The results of the experiment show that the GCNN model works better than other methods, with an accuracy rate of 89% on the chosen dataset. The study also expands its reach by looking at other “transfer learning (TL)” models, such InceptionV3 and Xception, which have accuracy rates of above 92%. This study helps with early DR identification and treatment by giving doctors an accurate and quick way to diagnose patients automatically. As an expansion, the project suggests creating a front-end interface that is easy to use using the Flask framework and adding authentication for safe user testing.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_17348112 |
| institution | Zenodo |
| language | eng |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Classification of Diabetic Retinopathy Disease Levels by Extracting Topological Features Using Graph Neural Networks UtilitasMathematica Diabetic retinopathy, graph neural networks, variational auto encoders, retinal image classification <p>Diabetic retinopathy (DR) is one of the most common causes of blindness in the world”. It has to be diagnosed quickly and accurately so that treatment may begin as soon as possible. Manual examination of fundus photographs by doctors is prone to mistakes and is labor-intensive. “Using computer-assisted methods, especially Convolutional Neural Networks (CNNs), to automate DR diagnosis seems promising”. The goal of this research is to improve the processing of retinal pictures by using a “Graph Convolutional Neural network (GCNN)” to classify the severity of diseases. GCNNs increase feature extraction by using topological correlations in pictures. This makes classification results more accurate. The suggested GCNN model works well, as shown by evaluation measures including “accuracy, precision, recall, and F1-score”. The results of the experiment show that the GCNN model works better than other methods, with an accuracy rate of 89% on the chosen dataset. The study also expands its reach by looking at other “transfer learning (TL)” models, such InceptionV3 and Xception, which have accuracy rates of above 92%. This study helps with early DR identification and treatment by giving doctors an accurate and quick way to diagnose patients automatically. As an expansion, the project suggests creating a front-end interface that is easy to use using the Flask framework and adding authentication for safe user testing.</p> |
| title | Classification of Diabetic Retinopathy Disease Levels by Extracting Topological Features Using Graph Neural Networks |
| topic | Diabetic retinopathy, graph neural networks, variational auto encoders, retinal image classification |
| url | https://doi.org/10.5281/zenodo.17348112 |