Face & Emotion Recognition with TensorFlow and OpenCV

This project is a from-scratch implementation of a basic Face Recognition and Emotion Detection system using Python, TensorFlow, and OpenCV.

---

Features

• Custom Face Recognition: Train the model on your own dataset of faces (no pre-downloaded datasets).
• Emotion Detection: Classify faces into emotions such as Happy, Sad, Neutral, and Surprised.
• Webcam Integration: Capture and label training images directly from your webcam.
• Lightweight Model: Small and fast to train, focused on learning the fundamentals.
• Educational Focus: Demonstrates the machine learning pipeline rather than competing with production-grade systems.

---

Getting Started

1. Clone the Repository

2. Create a Virtual Environment

python -m venv .venv
source .venv/bin/activate   # Mac/Linux
.venv\Scripts\activate      # Windows

3. Install Dependencies

pip install -r requirements.txt

Dependencies:

1. tensorflow
2. opencv-python
3. numpy
4. matplotlib

4. Capture Training Data

Use the webcam capture script to collect images for each person/emotion. (Note: the data/ folder is git-ignored, so your personal images won’t be uploaded.)

Example:

python capture.py --type person --label NAME --count 200
python capture.py --type emotion --label EMOTION --count 200

5. Train the Model

python train_face_recognizer.py
python train_emotion_cnn.py

6. Run Real-Time Detection

python realtime_demo.py

Project Structure

face-emotion-project/
├─ capture.py
├─ train_face_recognizer.py
├─ train_emotion_cnn.py
├─ realtime_demo.py
├─ requirements.txt
├─ models/
└─ data/
   ├─ faces/            # for face recognition: data/faces/<person_name>/<img>.jpg
   └─ emotions/         # for emotion training: data/emotions/<emotion_label>/<img>.jpg