Smile Detection using MediaPipe and OpenCV

This project detects whether a person is smiling or not using MediaPipe and OpenCV.
It works by tracking facial landmarks around the lips and analyzing their positions in real time.

---

How it Works

1. The webcam captures video frames using OpenCV.
2. MediaPipe’s face landmark model detects 468 facial landmarks.
3. The program uses four key landmarks:
   • 61: left lip corner
   • 291: right lip corner
   • 13: upper lip center
   • 14: lower lip center
4. It calculates:
   • Mouth width: distance between the left and right lip corners
   • Mouth height: distance between the upper and lower lips
5. The ratio of width to height determines if the person is smiling.

If the ratio is above a threshold (around 1.8), the system labels the person as smiling. Otherwise, it labels them as neutral.

---

Requirements

• Python 3.8 or higher
• OpenCV
• MediaPipe
• NumPy

Install dependencies with:

pip install opencv-python mediapipe numpy

---

How to Run

1. Download the MediaPipe model file face_landmarker.task from the MediaPipe documentation.
2. Place it in the same directory as the Python scripts.
3. Run the main script:

python smile_detection.py

4. Press q to quit the program.

---

Project Structure

smile-detection/
├── smile_detection.py        # Main smile detection script
├── detector.py               # Initializes and draws landmarks
├── face_landmarker.task      # MediaPipe model file
└── README.md                 # Project documentation

---

Customization

You can adjust the smile detection threshold in smile_detection.py:

if smile_ratio > 1.8:

Lower values make it easier to detect a smile. Higher values make detection stricter.

---