emotion-detection-media-pipe

🎭 Emotion Detection using MediaPipe & OpenCV

This project detects human emotions in real-time using MediaPipe FaceMesh and OpenCV.
It analyzes facial landmarks to classify simple emotions such as Happy, Sad, Surprised, and Neutral.

---

🧠 Project Overview

This project uses:

• MediaPipe FaceMesh for facial landmark detection (eyes, lips, etc.)
• OpenCV for image/video frame processing
• A custom rule-based classifier to determine emotion from facial features

It runs in Google Colab and saves the output video with detected emotions displayed on-screen in bold text.

---

🚀 Features

✅ Detects facial emotions in video files
✅ Uses real-time FaceMesh landmark extraction
✅ Displays the emotion name on screen in large, dark text
✅ Saves the processed output video to Google Drive
✅ Beginner-friendly — easy to extend with ML-based classifiers

---

📂 Folder Structure

Emotion-Detection-MediaPipe/
│
├── emotion_detection.ipynb     # Main Google Colab notebook
├── input_videos/               # Folder containing input videos
├── output_videos/              # Folder where processed videos are saved
└── README.md                   # Project documentation

---

🧩 Dependencies

Install the required libraries before running the notebook:

pip install opencv-python mediapipe numpy scipy

---

🧾 How It Works

1. Load video from Google Drive
2. Detect facial landmarks using MediaPipe FaceMesh
3. Compute geometric ratios (eyes, mouth, eyebrows)
4. Classify emotion based on feature distances
5. Overlay detected emotion text on the frame
6. Save the processed video with emotion labels

---

💡 Emotion Logic (Rule-Based)

---

🎥 Example Output

When the model processes a video, you’ll see text such as:

Emotion: Happy

<img width="552" height="887" alt="image" src="https://github.com/user-attachments/assets/6eff93fb-b132-4362-bb37-379d35506b2a" />

displayed clearly on the person’s face in large, dark text.

---

📁 Example Paths

Make sure to update the paths in your notebook before running:

video_path = "/content/drive/MyDrive/AI ML Projects/Emotion-Detection/input/video1.mp4"
output_path = "/content/drive/MyDrive/AI ML Projects/Emotion-Detection/output/video1_output.mp4"

---

🧪 Future Work

• Add a CNN-based deep learning emotion classifier
• Real-time webcam emotion recognition
• Integration with sentiment-based alert systems
• Emotion analytics dashboard for study or HR tools

---

🧑‍💻 Author

Fatima Noor
AI/ML Developer | Python Enthusiast
📧 [GitHub Profile Link Here]
📍 Based in Pakistan

---

🪄 Inspiration

The project was inspired by how subtle changes in facial expressions reveal human emotions — bringing AI closer to understanding empathy.
Built with ❤️ using MediaPipe, OpenCV, and Python.