🏥 VedaralaAI - Intelligent Skin Health Platform

VedaralaAI is a full-stack healthcare platform that leverages AI to analyze skin conditions, explain their severity, and assist patients in finding nearby dermatologists.

📌 Key Features

• 🧠 AI-Powered Skin Analysis using EfficientNetB0 + Tensor Cores
• ⚖️ Severity Classification (Mild, Moderate, Severe)
• 🧬 Detailed Disease Explanation via local JSON (offline-capable)
• 🌍 Doctor Locator using Google Maps API
• 🔐 Secure User Authentication (Register/Login)
• 📜 Patient History with Saved Reports
• 📄 PDF Export of Analysis Results
• 📷 Live Camera Scan + Image Upload
• 🌐 Modern Responsive UI using React

🧱 Project Stack

🔧 System Requirements

📁 Folder Structure

📂 Backend (/backend)

📂 Frontend (/frontend/src)

📂 Model (/model-2)

⚙️ Environment Variables (.env)

Place this in your /backend folder as .env:

GOOGLE_API_KEY=your_google_api_key
GOOGLE_MAPS_EMBED_KEY=your_embed_key

📄 Backend Setup

cd backend
pip install -r requirements.txt
python setup_database.py
python app.py

📌 Runs on http://localhost:5000

💻 Frontend Setup

cd frontend
npm install
npm start

📌 React app on http://localhost:3000

🧪 Default Users

📦 Backend Dependencies

Here's what's included in requirements.txt:

flask==2.3.3
flask-cors==4.0.0
mysql-connector-python==8.1.0
bcrypt==4.0.1
numpy==1.24.3
tensorflow==2.10.0
Pillow==10.0.0
python-dotenv==1.0.0

📦 Frontend Dependencies (via npm)

npm install jspdf html2canvas

These are used for exporting reports as PDF from the browser.

🧠 AI Model

• EfficientNetB0 with custom classifier
• Input: 224x224 RGB images
• Output: Benign, Melanoma, Nevus
• Severity estimation based on confidence thresholds
• Explanation fetched from disease_explanations.json

🌍 Google Maps Integration

• GET /api/get-map-key - Retrieves frontend embed key
• POST /api/search-doctors - Searches dermatologists near your geolocation

🧠 Prediction + Explanation Flow

1. Upload/capture image in Scan Page
2. /api/predict returns prediction, confidence, severity
3. /api/explain-disease returns matching explanation
4. Frontend renders animated explanation + nearby dermatologists

🛡️ Security Notes

• ✅ Passwords hashed with bcrypt
• ✅ Sessions managed via Flask with secure cookies
• ✅ Secret key stored in app.secret_key (change it in production)
• ✅ Uses .env to store sensitive keys

📊 Database Tables

users

Stores login credentials

skin_analysis

Stores scan results, severity, image path

patient_records

(Placeholder) for future patient medical history

📄 Report PDF Export

• ScanPage.jsx uses html2canvas + jsPDF
• User can download diagnosis as a styled PDF
• Includes image, prediction, severity, timestamp

🧪 Testing the Model

python test_model.py

Optional CLI test:

python predict.py path/to/image.jpg

✨ Future Enhancements

• 👨‍⚕️ Doctor profiles + booking
• 🧾 Report export to email / cloud
• 📈 Model performance dashboard
• 🗂️ Multi-disease classification

📄 License

MIT License – Free to use and modify
© 2025 VedaralaAI

📞 Support

• 📧 Email: support@vedaralaai.com
• 🐛 Issues: GitHub Issue Tracker
• 📚 Docs: This README.md

📽️ Video Demo

🔗 Prefer watching instead of downloading?
Click here to stream the demo on Google Drive