StayAwake-AI

An AI-powered application designed to detect driver drowsiness in real-time, enhancing road safety by alerting drivers to potential fatigue.

Project Structure

StayAwake-AI/
├── .devcontainer/                    # Development container configuration
├── diagrams/                         # Project diagrams and visualizations
├── docs/                            # Documentation files
│   ├── efficientnet_architecture_comparison.md
│   └── relevant data sets and resources.md
├── models_inattention/              # Model results and performance metrics
│   ├── B0/                         # EfficientNet-B0 results
│   ├── B1/                         # EfficientNet-B1 results
│   ├── B2/                         # EfficientNet-B2 results
│   ├── B4/                         # EfficientNet-B4 results
│   └── ImageNet_Pretrained_B4/     # Pre-trained B4 results
├── notebooks/                       # Jupyter notebooks
│   ├── driver-drowsiness-detection-case-study.ipynb
│   └── README.md
├── only_cm/                        # Confusion matrices and evaluation plots
├── only_models/                    # Trained model weights (.keras files)
├── scripts/                        # Utility scripts
│   ├── fetch_learning_rates.py
│   └── models_diagrams.py
├── model_card.md                   # Comprehensive model documentation
├── run notebook.md                 # Instructions for running notebooks
├── requirements.txt                # Python dependencies
└── README.md                      # Project documentation

How to Run the Project

🚀 Recommended: Run on Kaggle (Easiest)

Option 1: View and Run Directly

1. Visit the live notebook: Driver Drowsiness Detection on Kaggle
2. View the results: See all outputs, visualizations, and model performance
3. Run immediately: Click "Copy and Edit" to create your own editable version

Option 2: Start from Template

For the best experience with pre-configured environment and access to all datasets and models:

1. Visit the Kaggle Notebook: Access the complete notebook with all resources
2. Click "Copy and Edit": Create your own version of the notebook
3. Enable GPU (recommended): Go to Settings → Accelerator → GPU
4. Run All Cells: Execute the entire notebook with one click

What's included on Kaggle:

• ✅ Complete driver inattention dataset
• ✅ Pre-trained EfficientNet models (B0, B1, B4)
• ✅ All required Python libraries
• ✅ Sample images for testing
• ✅ Visualization and evaluation code

🛠️ Alternative: Local Development

Prerequisites

• Python 3.8+
• GPU with CUDA support (recommended)
• Git

Setup Instructions

1. Clone the repository:

git clone https://github.com/veyselserifoglu/StayAwake-AI.git
cd StayAwake-AI

2. Install dependencies:

pip install -r requirements.txt

3. Download required resources:

   • Trained model weights
   • Training dataset
   • OOD dataset

4. Adjust file paths in the notebook to match your local directory structure

5. Run the notebook:

jupyter notebook notebooks/driver-drowsiness-detection-case-study.ipynb

Note: Local execution requires significant setup and may encounter path/dependency issues. Kaggle is strongly recommended for the best experience.

Development Setup (Advanced)

You can choose between two development environments:

Option 1: Local Development with VS Code Dev Container

Prerequisites

• Docker
• VS Code
• VS Code Remote - Containers extension

Getting Started

1. Clone the repository:

git clone https://github.com/veyselserifoglu/StayAwake-AI.git
cd StayAwake-AI

2. Open the project in VS Code:

code .

3. When prompted, click "Reopen in Container" to start the development environment.

4. The container will automatically:

   • Install all required dependencies
   • Set up the Python environment
   • Configure VS Code with recommended extensions

Option 2: Direct Local Setup

1. Follow the local development setup instructions above
2. Use your preferred IDE or text editor
3. Install Jupyter for notebook development:

pip install jupyter
jupyter notebook

Model Information

This project implements driver inattention detection using EfficientNet architectures:

• EfficientNet-B0: Lightweight model with 84.2% validation accuracy
• EfficientNet-B1: Balanced model with 86.1% validation accuracy
• EfficientNet-B4: High-performance model with 88.7% validation accuracy
• EfficientNet-B4 (ImageNet Pretrained): Best performing model with 91.3% validation accuracy

For detailed model information, see model_card.md.

Files Description

• notebooks/: Contains the main Jupyter notebook with complete implementation
• model_card.md: Comprehensive documentation of all models and their performance
• run notebook.md: Quick start guide for running the notebook
• models_inattention/: Performance metrics and results for each model variant
• only_models/: Pre-trained model weights in Keras format
• only_cm/: Confusion matrices and evaluation visualizations
• scripts/: Utility scripts for model analysis and visualization

Contributing

1. Create a new branch for your feature:

   git checkout -b feature/your-feature-name

   • Branch naming convention: feature/ for new features, fix/ for bug fixes
   • Each branch should focus on a single feature or fix
   • Branches should not live longer than one week
   • If work takes longer than a week, create a new branch and reference the previous work

2. Make your changes and commit them:

   git add .
   git commit -m "feat: description of your changes"

3. Push your branch and submit a pull request:

   git push origin feature/your-feature-name

   • Create a pull request from your branch to main
   • Include a clear description of your changes
   • Reference any related issues

4. After your pull request is merged:

   • Delete your feature branch
   • Update your local main branch

   git checkout main
   git pull origin main

License

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