DriveSense Wellness Monitor 🚗🧠

DriveSense is an AI-powered driver wellness monitoring system that detects emotional states and signs of distress in real-time. Using Computer Vision and TensorFlow Lite, it identifies when a driver might be anxious or distressed and automatically triggers calming music to help them relax.

✨ Features

• Real-time Emotion Detection: Analyzes facial expressions to identify "Distressed/Anxious" vs. "Happy/Calm" states.
• Automated Music Therapy: Seamlessly plays curated tracks from your collection when distress is detected.
• Premium HUD Overlay: A futuristic, semi-transparent interface (Glassmorphism) for clear status updates and "Now Playing" indicators.
• Quantized Model: Efficient INT8 TFLite model for optimized performance on edge devices.
• Dynamic Threshold Support: Toggle between sensitivity levels for different environments.

🚀 Getting Started

Prerequisites

• Python 3.8+
• Webcam
• Audio output device

Installation

1. Clone the repository:

   git clone https://github.com/AaryanGole26/DriveSense.git
   cd DriveSense

2. Set up the virtual environment:

   python -m venv venv
   source venv/Scripts/activate  # Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

Configuration

1. Place your MP3 tracks in the songs/ directory.
2. Ensure you have the distress_model.tflite file in the root directory.
3. Keep songs/ and dataset/ local on your machine (they are intentionally not pushed to git).

🎮 Controls

Key	Action
Q	Quit Program
P	Pause/Resume Music
T	Toggle Sensitivity (0.5 ↔ 0.7)
V	Toggle Verbose Console Logs
F	Toggle Fullscreen Mode
(X)	Close Window to Exit

🛠 Project Structure

• wellness_monitor.py: Main application entry point.
• train_quantized_model.py: Training script for the TFLite model.
• prepare_dataset.py: Utility for dataset preprocessing.
• songs/: Directory containing calming tracks.
• dataset/: Training data source.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.