Injury Prediction & Prevention App (Android + CNN)

Project Objective

Deliver an intuitive, privacy-aware first aid android app that helps users quickly assess common injuries and find credible next steps. The solution blends computer vision with rule-based reasoning and location services to move users from confusion to action in minutes.

How It Works

1. Capture: User selects/takes an injury photo (cuts, bruises, burns, bleeding).
2. Questionnaire: Short form captures pain level, duration, incident cause, age group, allergies, etc.
3. CNN Inference: The photo is normalized and run through a TFLite CNN to predict injury class with confidence.
4. Decision Tree: Form responses are scored to estimate severity and flag red-alerts (e.g., heavy bleeding, burns on face/hands).
5. Guidance: The app shows first-aid steps, a link to nearby hospitals on Google Maps, and relevant video tutorials.
6. Chatbot: An AI assistant answers follow-up questions in simple language.

Project Modules

Key Features & Benefits

• Fast triage—combine visual clues with context for better suggestions.
• Explainable results—confidence scores and rule highlights.
• Nearby care—maps, distance, and one-tap calling.
• Education-first—trusted first-aid tutorials in a dedicated tab.
• Secure & private—authentication and minimal data retention.

Android Integration Sketch (Java/XML)

// PSEUDO-CODE (illustrative only; no Firebase JSON)

// 1) Image capture + CNN (TFLite)
class CnnClassifier {
  Interpreter tflite;
  List labels;
  CnnClassifier(AssetManager am){
    // load model.tflite and labels.txt
  }
  Result predict(Bitmap bmp){
    // resize -> float32 tensor -> tflite.run() -> softmax
    // return {className, confidence}
  }
}

// 2) Questionnaire + Decision Tree
Severity assessSeverity(Answers a){
  // Example rules:
  // if a.bleeding == "heavy" or a.burnArea > threshold -> HIGH
  // else if a.pain >= 7 and swelling == true -> MEDIUM
  // else -> LOW
  // return severity + reasons
}

// 3) Combine outputs
Assessment combine(Result vision, Severity sev){
  // Merge class + severity -> recommendation + red-alerts
}

// 4) Nearby hospitals (Maps)
void showNearbyHospitals(double lat, double lng){
  // Query Places API or pass coordinates to Google Maps intent:
  // Uri uri = Uri.parse("geo:" + lat + "," + lng + "?q=hospitals");
  // startActivity(new Intent(Intent.ACTION_VIEW, uri));
}

// 5) Tutorials
void openTutorial(String ytId){
  // startActivity(Intent.ACTION_VIEW, Uri.parse("https://www.youtube.com/watch?v=" + ytId));
}

// 6) AI Chatbot (edge/callable)
ChatReply askBot(String question, UserContext ctx){
  // sanitize -> call endpoint -> return short guidance
}
              

What You Get