Emotion Tracker

A professional Python desktop application for real-time facial emotion detection using DeepFace. The application analyzes user emotions via webcam and stores data as time-series CSV files with comprehensive analytics.

🎥 Demo Video

Watch the demo GIF above

Features

• Real-time Emotion Detection: Uses DeepFace for accurate emotion recognition
• Interactive GUI: Three-tab interface (Control, Monitor, Analytics)
• Performance Optimized: Processes every 3rd frame to reduce CPU load
• Data Storage: Automatically saves daily CSV files with timestamps
• Visual Analytics: Pie charts and statistics for emotion analysis
• Privacy-Focused: All data stored locally, no internet transmission

Requirements

• Python 3.11+ (tested with Python 3.11.1)
• Webcam access
• Internet connection (for initial DeepFace model download)
• macOS 11.0+ / Windows 10+ / Ubuntu 18.04+

Installation

1. Clone this repository:

   git clone https://github.com/DeepPythonist/EmotionTracker.git
   cd EmotionTracker

2. Create virtual environment:

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

3. Install dependencies:

   pip install -r requirements.txt

4. Run the application:

   python main.py

Usage

Control Tab

• Click START to begin emotion detection
• Click STOP to stop recording
• Monitor status and frame count

Monitor Tab

• View current detected emotion in large text
• See confidence level with progress bar
• Track session statistics and top emotions

Analytics Tab

• Select date from dropdown to view historical data
• View emotion distribution pie chart (if matplotlib available)
• Review summary statistics
• Monitor current session emotions

Data Storage

• CSV files are automatically created in the data/ directory
• File naming format: emotions_YYYY-MM-DD.csv
• Columns: timestamp,emotion,confidence
• Data is appended continuously during recording sessions

Performance Optimization

The application implements several optimization techniques:

• Processes every 3rd frame (not every frame)
• Uses enforce_detection=False for better edge case handling
• Implements frame skipping when processing takes >100ms
• Resizes frames to 640x480 before processing
• Caches DeepFace model on first load
• Uses threading to separate UI from processing

Error Handling

• Camera not found: Shows error dialog, disables START button
• DeepFace import error: Shows installation instructions
• No face detected: Skips frame silently
• Low confidence (<0.5): Marks as "uncertain"

Project Structure

EmoNet/
├── main.py                    # Main application
├── requirements.txt           # Dependencies
├── README.txt                # Installation guide
├── PRIVACY.txt               # Privacy policy
├── LICENSE.txt               # MIT License
├── core/                     # Core modules
│   ├── camera_handler.py
│   ├── emotion_detector.py
│   └── data_manager.py
├── ui/                       # UI components
│   ├── main_window.py
│   ├── monitor_tab.py
│   └── analytics_tab.py
├── utils/                    # Utility modules
│   └── config.py
└── data/                     # Data storage folder

Technical Details

• CPU Usage: <15% on modern i5 processors
• RAM Usage: <500MB
• Detection Latency: <2 seconds per processed frame
• UI Responsiveness: No freezing during processing

Dependencies

• deepface==0.0.79
• opencv-python==4.8.1.78
• pandas>=2.0.0,<2.2.0
• matplotlib>=3.7.0,<3.9.0
• seaborn>=0.12.0,<0.14.0
• pillow>=9.0.0,<11.0.0
• tensorflow>=2.13.0,<2.16.0
• numpy>=1.21.0,<2.0.0

Support

If you have any issues:

• Email: mrasolesfandiari@gmail.com
• GitHub: https://github.com/DeepPythonist/EmotionTracker

License

MIT License - See LICENSE.txt for details.

This project is provided for educational and research purposes.