MediRoot is a mobile health assistant that combines OCR technology and NLP-powered symptom analysis to provide instant medical insights from prescriptions, reports, or user-described symptoms.
- 📸 Medical Text Extraction** - Scan medicine labels/reports using OCR (Gemini API)
- 💬 Symptom Analysis Chatbot** - Get AI-powered health guidance
- 🔊 Voice Output** - Hear diagnoses & recommendations via TTS
- 📊 Health Literacy Tools** - Simplified medical information
| Component | Technologies |
|---|---|
| Mobile App | Kotlin, Android SDK |
| Backend | Python, Flask |
| AI/ML | Gemini API, NLTK, NLP |
| OCR | Google Vision API |
| TTS | Android Text-to-Speech |
- Android Studio (for development)
- Python 3.8+ (for backend)
- Google Cloud API key (for Gemini/Vision)
# Clone repo
git clone https://github.com/ya5h8/MediRoot.git
# Android App
cd Android_App
Open in Android Studio
# Backend
cd Backend
pip install -r requirements.txt-
Tap "Upload Image" button
- Only accesses device gallery (no camera permission needed)
-
Select medicine image
- Choose clear photos of:
✓ Medicine labels
✓ Prescription slips
✓ Medical reports
- Choose clear photos of:
-
Press "Submit"
- Processes image via Gemini OCR
- Typical extraction time: 3-5 seconds
-
View Results
- Extracted text appears in this format:
[Medicine Name]: Paracetamol [Dosage]: 500mg every 6 hours [Warnings]: Do not exceed 4g/day
- Extracted text appears in this format:
-
Type symptoms in text box
(Example: "headache and fever since yesterday") -
Hit "Enter" to get:
- 🩺 Possible Condition
- 💊 Suggested Remedies
⚠️ When to See Doctor
- Tap "Speak" to hear:
- Extracted medicine details
- Diagnosis explanations
📌 Best Practices:
- Use well-lit, high-resolution images
- Crop to focus on relevant text
- Supported formats: JPG, PNG (max 5MB)