Health AI Platform

Features

• Skin cancer detection using a trained Keras model
• Mental health assessment and recommendations
• Resource API for short videos (YouTube integration for skin cancer and mental health)

---

Skin Cancer Detection API

Model Setup

1. Export your trained Keras model as skin_cancer_model.keras.
2. Create a folder named ml-models at the root of your project.
3. Place your model file at:

   ml-models/skin_cancer_model.keras
   

Backend API

• Endpoint: POST /api/skin-assessment/detect
• Request: Send a multipart/form-data POST request with an image file under the field name image.
• Response:

  {
    "prediction": "benign", // or "malignant"
    "confidence": 0.87
  }

• Example cURL:

  curl -X POST http://localhost:3000/api/skin-assessment/detect \
    -F "image=@/path/to/your/image.jpg"

---

YouTube Resource API

• Endpoint: GET /api/resources
• Returns a list of recent public YouTube videos about skin cancer or mental health (auto-fetched using the YouTube Data API).
• To use, set your YouTube Data API key in pages/api/resources/index.ts.
• Example response:

  [
    {
      "id": "QZCGhY5tZ0w",
      "title": "Skin Cancer Awareness",
      "description": "A short YouTube video about skin cancer signs and prevention.",
      "url": "https://www.youtube.com/watch?v=QZCGhY5tZ0w",
      "category": "skin-cancer",
      "tags": [],
      "thumbnail": "https://img.youtube.com/vi/QZCGhY5tZ0w/0.jpg",
      "author": "Skin Health Channel"
    }
  ]

---

Dependencies

• @tensorflow/tfjs-node
• sharp
• formidable
• node-fetch (for YouTube API)

---

Setup

1. Install dependencies:

   npm install

2. Add your Keras model as described above.
3. (Optional) Add your YouTube Data API key for dynamic video resources.
4. Start the development server:

   npm run dev

---

License

MIT