Aravind Eye Hospital in India hopes to detect and prevent diabetic retinopathy[1] among people living in rural areas where medical screening is difficult to conduct[2]. Diabetic retinopathy (DR) is a leading cause of blindness and affects up to 80 percent of those who have had diabetes for 20 years or more [3].

The data is collected from multiple clinics in Indian rural areas using fundus photography over an extended period of DR. The data source is from a Kaggle competition [4]. The public dataset size is 10G, consisting of a pre-split training set (3662 images and 1 outcome label) and a test set (1928 images). The private test set consisting of 20GB of data across 13,000 images. The outcome label is the severity of diabetic retinopathy rated by clinicians on a scale of 0 to 4.

The goal of the project is to train an image classifier to automatically identify diabetic retinopathy and provide information on the severity of the disease. We will follow a machine learning pipeline to conduct analysis:

1. Data Overview

2. Data Preprocessing

3. Data Augmentation

4. Modeling

5. Result

[1] Retrieved from: https://nei.nih.gov/learn-about-eye-health/eye-conditions-and-diseases/diabetic-retinopathy
[2] Retrieved from: https://www.kaggle.com/c/aptos2019-blindness-detection/overview
[3] Retrieved from: https://www.kaggle.com/tanlikesmath/intro-aptos-diabetic-retinopathy-eda-starter.
[4] Retrieved from: https://www.kaggle.com/c/aptos2019-blindness-detection/data
[5] Retrieved from: https://www.kaggle.com/aleksandradeis/aptos2019-blindness-detection-eda