Government doctor per 11,000 people in India compared to 1 doctor per 400 people in the USA. India has 17% of the global population and 20% of the global disease burden. Heart diseases and pneumonia are among the top causes of death. Diabetes mellitus is directly linked to cardiovascular diseases with an estimated 2 million deaths per year.70% of practitioners in rural India have no formal training / are quacks. Users should be able to predict and monitor common diseases at home. Users should have access to accurate diagnostic information. Users should be able to browse, extract and digest the information in a user-friendly manner through the product in their regional language. Even illiterate people will be able to use this app, All they need to do is upload the pictures of their report. The Tesseract OCR recognization helps to auto input the fine details of the reports. Further Machine Learning model will be able to correctly interpret the diseases and also suggest the required tests and advice when to visit the nearest hospital at a very early stage via a text message sent to the registered phone number. Our solution also keeps track of the last 10 sugar levels and suggests the necessary tests and hospital visits. Our Machine Learning models are able to correctly interpret Pneumonia via X-ray Images, Diabetes via insulin and Glucose levels, Malaria via Bite mark Images and blood smear.
Extracting the details from the pictures of reports using Tesseract OCR recognization took most of our time. Sending the Text message to the registered phone was the one thing we had never done before, so we needed to learn that particular functionality. Suggesting the nearest hospital to the person to get the actual Medical consultation was one facility we were not able to implement due to lack of time.
Discussion