हेल्थCare
A virtual AI-based assistant that displays the user's health status using lab reports. It also consists of an AI-based chatbot that is useful for people to ask queries.
Created on 2nd June 2023
•
हेल्थCare
A virtual AI-based assistant that displays the user's health status using lab reports. It also consists of an AI-based chatbot that is useful for people to ask queries.
The problem हेल्थCare solves
This AI based model can be helpful for the assistant present in the laboratories while the doctors are not available or are out of reach due to any reason . This can be helpful in determining the health issues of the patient during its visit to labs . This is a trained model so doctor's consultancy is a must but still many a times there can be possiblities of emergency situations and hence this model can be helpful at that instance .
This website can not only be used by just assistant but can also be used by laymann people because the web-app consist of extra features such as a AI based chatbot which can be helpful for the user to ask there queries and get resolve with best possible AI outcomes . Also , it have consist of a map which use users latitude and longitude to fetch current location and then zoom out to 5KM radius of the user displaying all nearby hospitals . So , they can rush to nearby hospital as soon as possible in case of any emergency situation .
Challenges we ran into
1} Model Training : We faced a lot of challenges in training our model for this project . While training EYE model the data set was quite large and the initial accuracy was not that accurate so we have use image processing and tried to increase the accuracy upto 98%.
2} model deployment : We faced difficulties in deploying the model as we were using MERN stack for our frontend and backend . But we used tensorflow.js to resolve all our queries .
3} chatbot : We faced a lot of problems while displaying the response of chatbot from backend to frontend as we messed up in 'cors' . yet we inspected console / network and preview by the advice of the mentors and were successful to integrate backend and frontend .
4} deployment : we are unable to deploy the website because of large dataset present in the backend and limit being just 50MB