Emoji Reactor

A real-time camera-based emoji display application that uses MediaPipe to detect your poses and facial expressions, then displays corresponding emojis in a separate window.

Features

• Hand Detection: Raises hands above shoulders → displays hands up emoji 🙌
• Smile Detection: Detects smiling → displays smiling emoji 😊
• Default State: Straight face → displays neutral emoji 😐
• Real-time Processing: Live camera feed with instant emoji reactions

Recent Updates

• Added new emoji assets: jawline.png and 67clashroyale.webp.
  • These are bundled in the project root and can be used as alternative display images.
  • To use them, either replace one of the default image files or extend emoji_reactor.py to load additional states.

Requirements

• Python 3.12 (Homebrew: brew install python@3.12)
• macOS or Windows with a webcam
• Required Python packages (see requirements.txt)

Installation

1. Clone or download this project

2. Create a virtual environment (Python 3.12) and install deps:

   # macOS: ensure Python 3.12 is installed
   brew install python@3.12
   
   # Create and activate a virtual environment
   python3.12 -m venv emoji_env
   source emoji_env/bin/activate
   
   # Install dependencies
   pip install -r requirements.txt

3. Ensure you have the emoji images in the project directory:

   • smile.jpg - Smiling face emoji
   • plain.png - Straight face emoji
   • air.jpg - Hands up emoji
   • jawline.png - Jawline-themed emoji asset
   • 67clashroyale.webp - 67 Clash Royale-themed emoji asset

Usage

1. Run the application:

   # Option A: use helper script
   ./run.sh
   
   # Option B: run manually
   source emoji_env/bin/activate
   python emoji_reactor.py

2. Two windows will open:

   • Camera Feed: Shows your live camera with detection status
   • Emoji Output: Displays the corresponding emoji based on your actions

3. Controls:

   • Press q to quit the application
   • Raise your hands above your shoulders for hands up emoji
   • Smile for the smiling emoji
   • Keep a straight face for the neutral emoji

How It Works

The application uses two MediaPipe solutions:

1. Pose Detection: Monitors shoulder and wrist positions to detect raised hands
2. Face Mesh Detection: Analyzes mouth shape to detect smiling vs. straight face

Detection Priority

1. Hands Up (highest priority) - Overrides facial expression detection
2. Smiling - Detected when mouth aspect ratio exceeds threshold
3. Straight Face - Default state when no smile is detected
4. Finger on Jawline - Detected when you flex your jawline with your finger pouch
5. 67 Emote - Detected when you move your hands like the 67 meme

Customization

Adjusting Smile Sensitivity

Edit the SMILE_THRESHOLD value in emoji_reactor.py:

• Decrease value (e.g., 0.30) if smiles aren't detected
• Increase value (e.g., 0.40) if false positive smiles occur

Changing Emoji Images

Replace the image files with your own:

• smile.jpg - Your smiling emoji
• plain.png - Your neutral emoji
• air.jpg - Your hands up emoji
• jawline.png - Flex your jawline (if you have)
• 67clashroyale.webp - Your 67 meme

Troubleshooting

Camera Issues (macOS)

• If you see "not authorized to capture video", grant Camera access for your terminal/editor:
  • System Settings → Privacy & Security → Camera → enable for Terminal/VS Code/iTerm
• Quit and relaunch the terminal/editor after changing permissions
• Ensure no other app is using the camera
• Try different camera indices by changing cv2.VideoCapture(0) to cv2.VideoCapture(1)

Detection Issues

• Ensure good lighting on your face
• Keep your face clearly visible in the camera
• Adjust SMILE_THRESHOLD if needed
• For hands up detection, make sure your arms are clearly visible

Technical Details

• Uses OpenCV for camera capture and display
• MediaPipe Pose and FaceMesh for detection
• Real-time RGB conversion and landmark detection

Dependencies

• opencv-python - Computer vision library
• mediapipe - Pose and Face Mesh detection
• numpy - Numerical computing

See requirements.txt for installation and requirements-lock.txt for pinned versions.

License

MIT License.