HealthVision is a MedTech web application built using the MERN stack (MongoDB, Express.js, React, Node.js) with integrated AI/ML capabilities.
The project uses Git-based branching workflows for feature development and version control.
Provide users a simple, accessible digital health platform where they can analyze medical reports, book appointments, and consult doctors, all through a unified and secure dashboard.
- Secure login & signup flow using JWT + bcrypt
- Access to dashboard and all personalized features only when logged in
- Browse list of doctors with basic details
- View patient feedback & ratings
A personalized dashboard containing:
- Quick navigation tiles for key actions
- Simple charts representing user activity (appointments history, consultations, etc.)
- Upload a medical report (PDF/Image)
- Extract text using OCR
- Pass extracted data to lightweight ML models for basic disease pattern detection such as dengue, typhoid, viral fever
- Show simple result summary (not full diagnostic)
- Symptom Chatbot: Users can describe symptoms in a chatbot interface, which will analyze the input and suggest the urgency level for consulting a doctor (e.g., immediate, within 24 hours, routine consultation)
- Allow user to initiate voice/video consultation with doctor
- Basic real-time connection using WebRTC (no call recording, no complex room system yet)
Users can:
- Select doctor
- Pick timeslot
- Confirm appointment
- Allow user to pay 20% of consultation fee during booking
- Use a test-mode gateway (Razorpay/Stripe) for MVP
- Frontend: React.js
- Backend: Node.js + Express.js
- Database: MongoDB
- AI/ML: OCR + basic disease-prediction models (Python model served through API)
- WebRTC: For real-time communication
- Version Control: Git with feature-branch workflow
- Each feature (auth, report analysis, symptom chatbot, appointment, payment, WebRTC, dashboard) developed in a dedicated branch
- Merged into main via pull requests after testing
The HealthVision ecosystem relies on a 3-tier communication architecture between the Frontend, Backend, and AI/ML services:
-
Frontend (React) to Backend (Node.js):
- React sends HTTP requests (via Axios) to the Node.js REST API.
- Handles user actions like uploading report files (multipart/form-data) or fetching dashboard data.
- Secured via JWT tokens passed in the Authorization header.
-
Backend (Node.js) to AI/ML Service (Python/FastAPI):
- Node.js acts as an Orchestrator. When a report is uploaded, Node.js forwards the file or the file path to the Python ML API.
- Communication happens over internal HTTP requests (Service-to-Service).
- The Python service performs OCR and inference, returning structured JSON data (predicted diseases, confidence levels) back to Node.js.
-
Data Persistence (MongoDB):
- Node.js receives the ML results and persists them to MongoDB along with the user's report metadata.
-
Response Return:
- Node.js sends the final structured response back to the React frontend to be displayed on the analysis results page.
- Signaling: Node.js + Socket.io are used as a signaling server to exchange SDP (Session Description Protocol) and ICE candidates between two peers.
- Media Stream: Once signaling is complete, the peer-to-peer connection is established directly between the users' browsers for low-latency video/audio streaming.
