Diabetic Retinopathy Detection
Built an ensemble deep learning model combining VGG19, ResNet50, DenseNet121, and DenseNet169 architectures on the APTOS 2019 dataset. Achieved superior accuracy through class balancing, hyperparameter tuning, and comparative benchmarking across all model variants.
Diabetic Retinopathy Detection
About the Project
Built an ensemble deep learning model combining VGG19, ResNet50, DenseNet121, and DenseNet169 architectures on the APTOS 2019 dataset. Achieved superior accuracy through class balancing, hyperparameter tuning, and comparative benchmarking across all model variants.
Key Features
- Ensemble of VGG19, ResNet50, DenseNet121, DenseNet169
- APTOS 2019 dataset with class balancing
- Hyperparameter tuning & comparative benchmarking
- Automated 5-class retinopathy grading