A modern ML-powered biometric authentication system that identifies users based on their unique typing patterns. Your keystroke dynamics—how long you hold keys, the rhythm between keystrokes—create a behavioral biometric signature as unique as your fingerprint.

• 🔒 Behavioral Biometrics: Authenticate users based on how they type, not what they type
• 🤖 ML-Powered: Random Forest classifier with confidence scoring
• 📊 Real-time Analytics: Visualize typing patterns with interactive Plotly charts
• 🎨 Modern UI: Sleek dark-themed Streamlit interface
• ⚡ Fast Training: Quick model training with minimal samples

The system captures and analyzes three key typing characteristics:

flowchart LR
    subgraph capture [Keystroke Capture]
        A[Key Press] --> B[Key Release]
        B --> C[Timing Data]
    end
    
    subgraph features [Feature Extraction]
        D[Dwell Time]
        E[Flight Time]
        F[Digraph Latency]
        G[Trigraph Latency]
    end
    
    subgraph ml [ML Classification]
        H[Random Forest]
        I[User Prediction]
        J[Confidence Score]
    end
    
    C --> D & E & F & G
    D & E & F & G --> H
    H --> I & J

• Python 3.10 or higher
• pip package manager

# Clone the repository
git clone https://github.com/yourusername/keystroke-biometrics.git
cd keystroke-biometrics

# Create virtual environment (recommended)
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

streamlit run app.py

The app will open in your browser at http://localhost:8501

1. Navigate to the Register page
2. Enter a unique username
3. Type 3 sample paragraphs to train your profile
4. System learns your unique typing patterns

1. Navigate to the Authenticate page
2. Type the displayed paragraph naturally
3. Click "Authenticate" to identify yourself
4. View confidence score and probability distribution

1. Navigate to the Analytics page
2. Select a user to view their typing profile
3. Explore dwell times, digraph patterns, and feature importance

keystroke-biometrics/
├── app.py                 # Streamlit main application
├── src/
│   ├── __init__.py
│   ├── capture.py         # Keystroke timing capture
│   ├── features.py        # Feature extraction engine
│   ├── model.py           # ML model (Random Forest)
│   └── utils.py           # Helper functions
├── data/
│   └── users.json         # User keystroke profiles
├── models/
│   └── keystroke_model.joblib  # Trained model (generated)
├── requirements.txt
└── README.md

# Feature scaling for normalization
scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)

# Random Forest with balanced class weights
classifier = RandomForestClassifier(
    n_estimators=100,
    max_depth=10,
    class_weight='balanced',
    random_state=42
)

*Accuracy depends on number of users and sample quality

python -m pytest tests/

1. Add feature extraction in src/features.py
2. Update get_feature_names() and to_vector() methods
3. Retrain models with new feature set

Registration	Authentication

Sample Collection	Analytics Dashboard

• Keystroke Dynamics - Wikipedia
• Biometric Authentication via Keystroke Dynamics
• scikit-learn Random Forest

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

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

Charan

• GitHub: @charanpool

Made with ❤️ and ⌨️