Python OpenCV DeepFace TensorFlow Platform License

ANTRIX is a real-time AI-based emotion detection system built using Python, OpenCV, and DeepFace.
The system captures live webcam frames, analyzes facial expressions using deep learning, and predicts dominant human emotions in real time. ANTRIX is optimized for lightweight CPU-based execution and does not require dedicated GPU hardware. The project is designed with:

• real-time responsiveness
• modular architecture
• beginner-friendly code structure
• safe error handling
• cross-platform compatibility
• reproducible setup process

This project was developed as part of a Minor Project for Bachelor of Technology in Artificial Intelligence and Machine Learning.

• Real-time webcam-based emotion detection
• AI-powered facial emotion analysis
• Dynamic resizable OpenCV window
• Safe application termination
• Beginner-friendly modular code structure
• Cross-platform webcam handling
• Lightweight CPU-based execution
• Clean UI overlay with live information

ANTRIX can detect:

• Happy
• Sad
• Angry
• Fear
• Surprise
• Neutral
• Disgust

MinorProject2/
│
├── README.md
├── LICENSE
├── .gitignore
├── requirements.txt
├── main.py
│
├── Screenshots/
│ └── antrix-demo.png

• Intel i3 / Ryzen 3 or higher
• Minimum 4 GB RAM
• Webcam (Built-in or USB)
• 500 MB free storage

• Python 3.12
• pip package manager
• git
• Internet connection (first run only for model download)

git clone https://github.com/amartyasinghnamra/MinorProject2.git

Note:

If u dont have git installed run

winget install Microsoft.git

cd MinorProject2

python -m venv venv
venv\Scripts\activate

python3 -m venv venv
source venv/bin/activate

pip install -r requirements.txt

Run the application using:

python main.py

Note: The first launch may take longer because DeepFace downloads pre-trained model weights automatically.

1. Webcam initializes safely
2. Live frames are captured continuously
3. Frames are processed using DeepFace
4. Dominant emotion is detected
5. Emotion label is rendered on-screen
6. System loops in real time until exit

ANTRIX supports:

• Windows 10/11
• Linux
• macOS

The webcam initialization automatically adapts based on the operating system.

Possible fixes:

• Close Zoom/Teams/Discord
• Check camera permissions
• Restart system
• Try changing CAMERA_INDEX value

This is normal.

DeepFace downloads AI model weights during the first execution.

Upgrade pip before installation:

python -m pip install --upgrade pip

Then reinstall requirements.

Possible causes:

• Webcam already in use
• Invalid webcam index
• Camera permission denied

Potential future enhancements:

• Multi-face tracking
• Emotion history logging
• GUI interface using Tkinter/PyQt
• GPU acceleration
• Raspberry Pi deployment
• Mobile application integration
• Cloud synchronization

This project was developed for:

• academic learning
• computer vision practice
• AI experimentation
• affective computing research

• OpenCV
• DeepFace
• TensorFlow
• Python Community

Amartya Singh Namra

Bachelor of Technology
Artificial Intelligence and Machine Learning

This project is licensed under the MIT License.