-
-
Output of our project
Inspiration
1)Diabetic Retinopathy (DR) is a leading cause of blindness worldwide. It is caused by damage to the blood vessels in the retina, the light-sensitive tissue at the back of the eye. DR is a common complication of diabetes, affecting up to 50% of people with the condition over time. The rate of DR in India is estimated to be between 28% and 35%, which is significantly higher than the global average of 20%. 2)Diabetic readmission rates in India are estimated to be between 20% and 25%, which is significantly higher than the global average of 14.4–22.7%. This is likely due to a number of factors, including the high prevalence of diabetes in India, the lack of access to quality healthcare, and the poor awareness of diabetic complications among the general population.
What it does
Our project takes a retinal image as input and outputs the stage of diabetic retinopathy and the likelihood of readmission for patients who are admitted due to diabetic retinopathy. The readmission prediction input is a set of specific attributes, such as the patient's medical history, social history, and access to care.
How we built it
1)We took Diabetic Retinopathy images and readmission rate from kaggle. 2)We built a CNN model to classify the stages of Diabetic Retinopathy . 3)For readmission ,we have built a random forest classifier.
Challenges we ran into
1)Datasets were complex. 2)Limited references.
Accomplishments that we're proud of
In our project we introduced Diabetic Retinopathy detection with predicting Readmission rate of DR patients. So that a diabetes patient can detect Diabetic Retinopathy at early stages with the help of our model and Doctor's guidance. Also our model helps them to avoid readmission in case of surgery regarding Diabetic Retinopathy.
What we learned
We learnt how to deal with real time problems efficiently.
What's next for Detecting Diabetic Retinopathy & Readmission
Our future plan is to integrate this model with an App and also much quantization is required so that we can improve its accuracy ,So that it can operates in real time much efficiently.
Built With
- cnn
- keras
- python
- scikit-learn
- tensorflow
