This project focuses on the automated detection of retinal abnormalities by analyzing Optical Disc images. Using Deep Learning, the system classifies retina scans into two categories: Healthy and Unhealthy.
| Metric | Specification |
|---|---|
| Model Architecture | ResNet50V2 (Transfer Learning) |
| Accuracy | 98.7% |
| Frameworks | TensorFlow, Keras, OpenCV |
| Optimizer | Adam |
notebooks/: Comprehensive Jupyter Notebooks for training and testing.images/: Performance charts (Accuracy/Loss) and sample outputs.requirements.txt: Necessary dependencies for reproducibility.
Bu proje, retina görüntülerinden Optik Disk analizi yaparak göz sağlığı durumunu otomatik olarak sınıflandırmak amacıyla geliştirilmiştir. Derin Öğrenme yöntemleri kullanılarak, retina taramaları "Sağlıklı" ve "Sağlıksız" olarak sınıflandırılır. %98.7 başarı oranı ile tıbbi taramalar için güçlü bir karar destek mekanizması sunar.
- Model: ResNet50V2 mimarisi ile Transfer Learning (Görüntü İşleme).
- Veri Seti: Train/Test/Validation olarak ayrılmış tıbbi görüntü seti.
- Teknolojiler: Python, TensorFlow, Keras, OpenCV.
- Esma GELEBEK - Model Architecture & Development
- Esma Betül KOCAAHMET - Model Architecture & Development
- Funda BOZBURUN - Model Architecture & Development