Webapp for classification of Diabetic Retinopathy from retinal images using flask and keras.
Diabetic retinopathy is an eye condition that can cause vision loss and blindness in people who have diabetes. It affects blood vessels in the retina.
| No Diabetic Retinopathy | Severe Diabetic Retinopathy |
|---|---|
 |
 |
Dataset consists of retina images taken using fundus photography under a variety of imaging conditions.
A clinician has rated each image for the severity of diabetic retinopathy on a scale of 0 to 4:
0 - No DR
1 - Mild
2 - Moderate
3 - Severe
4 - Proliferative DR
Dense Convolutional Network (DenseNet), which connects each layer to every other layer in a feed-forward fashion. Whereas traditional convolutional networks with L layers have L connections - one between each layer and its subsequent layer.
A more efficient model variant DenseNet-BC (DenseNet-Bottleneck-Compressed) networks are trained. Using the DenseNet-BC-121-32 model as the base model.
Advantages of DenseNet-BC are:
- Reduced number of parameters
- Similar or Better performance
- Better accuracy
Dense Net architecture as shown in the original paper which shows the connections from each layer to every other layer:
virtualenv venv
source /bin/activate
python3 -m pip install -r requirements.txt
python3 -m pip install -r requirements.txt
python3 app.py