"Bridging the diagnostic gap — one scan at a time."
MediScan AI is a web-based AI diagnostic assistant that analyzes medical images — chest X-rays, skin photos, and retinal scans — and returns an instant diagnosis with a visual heatmap, urgency score, and plain-language explanation. Built for doctors and health workers in low-resource settings where radiologists are scarce.
- 2/3 of the world lacks access to diagnostic imaging
- Rural India has just 1 radiologist per 100,000 patients
- Average radiology wait time: 3–14 days
- Diseases like TB, pneumonia, and melanoma go undetected until it's too late
- 📸 Image Analysis — Chest X-rays, skin photos, retinal scans
- 🔥 Grad-CAM Heatmaps — Visually highlights the suspicious region
- 🎯 Credibility Score — AI confidence score (0–100)
- 🚨 Urgency Scoring — Low / Moderate / High / Critical
- 🌐 Multilingual — Results in 10+ languages
- 📱 Offline PWA — Works on 2G/3G, no install needed
- 🔐 HIPAA-Aligned — Images deleted post-analysis
| Layer | Technology |
|---|---|
| Frontend | React.js, Tailwind CSS, PWA |
| Backend | Python Flask, Node.js |
| AI Model | EfficientNet-B4, Grad-CAM |
| Database | PostgreSQL, Redis |
| Deployment | Docker, AWS EC2 |
# Clone the repo
git clone https://github.com/drop2life/mediscan-ai.git
cd mediscan-ai
# Frontend
cd frontend && npm install && npm start
# Backend
cd backend && pip install -r requirements.txt && python app.pymediscan-ai/
├── frontend/ # React.js UI
├── backend/ # Flask REST API
├── models/ # EfficientNet-B4 trained model
├── datasets/ # NIH + ISIC dataset references
└── docker-compose.yml
MediScan AI provides AI-based analysis and should not replace professional medical diagnosis. Always consult a qualified healthcare professional.
| Team Leader | Anmol Verma |
| aanmolverma1309@gmail.com | |
| College | ABES Engineering College |
Built with ❤️ by Team Drop2Life