face_emotion_detection

Face Emotion Detection

Small Flask + Socket.IO application that performs real-time face emotion detection using a prebuilt Keras model and OpenCV face detection. This project demonstrates streaming webcam frames to a browser, running emotion predictions on detected faces, and emitting live emotion data via WebSockets.

Overview

• Real-time video captured from a webcam is processed by OpenCV to detect faces.
• A prebuilt Keras model (src/emotion_model.h5) predicts emotion probabilities for each detected face.
• The app serves a video stream and emits JSON emotion updates via Socket.IO for dashboard visualization.

Prebuilt models & data

• Emotion classification model: src/emotion_model.h5 (pretrained Keras HDF5 model). Place this file in src/ (already included in this repo).

• Face detector: src/haarcascade_frontalface_default.xml (OpenCV Haar cascade). Also included in src/.

These are prebuilt assets bundled with the project; you can replace them with your own models if desired.

Project layout

• src/ — model and face cascade, and helper scripts.
• webapp/ — Flask application and frontend templates.
  • webapp/app.py — main Flask + Socket.IO server.
  • webapp/templates/ — HTML templates for home, login, dashboard.
  • webapp/static/ — CSS, JS and captures/ saved images.

Requirements

Python 3.8+ recommended. Key packages:

• flask
• flask-socketio
• tensorflow (or tensorflow-cpu depending on your environment)
• opencv-python
• numpy

I can add a requirements.txt for you — let me know.

Installation

1. Create and activate a virtual environment:

python -m venv .venv
.\.venv\Scripts\Activate.ps1

1. Install dependencies:

python -m pip install -r requirements.txt
# or if you don't have requirements.txt:
python -m pip install flask flask-socketio tensorflow opencv-python numpy

Run the app

From the webapp directory run:

python .\app.py

Open your browser at http://localhost:5000 and go to the dashboard to see the live video and emotion updates.

Notes:

• On Windows PowerShell you may need to run .\app.py or python app.py — python app.py is recommended.
• The app uses your default webcam (index 0). If you have multiple cameras, change the index in app.py where cv2.VideoCapture(0) is called.
• Socket.IO is used to broadcast emotion percentages in real time to clients.

Endpoints & features

• / — Home page
• /dashboard — Dashboard (requires a simple session login)
• /video_feed — MJPEG stream of the webcam frames
• /get_emotions — Returns latest emotion percentages as JSON
• /capture — POST endpoint that accepts a base64 image (image) and saves it to static/captures

Security & Production notes

• app.secret_key in webapp/app.py is a placeholder; change it before deploying.
• Running TensorFlow and webcam access in production requires careful resource handling; consider running model inference in a separate worker or using batching for scalability.

Troubleshooting

• Pylance warnings about missing packages typically mean your VS Code interpreter isn’t the one with your installed packages — select the correct Python interpreter in the bottom-right of VS Code.
• If the camera fails to open, ensure no other app is using it and try changing the device index.

Credits

This repository stitches together OpenCV face detection and a pretrained Keras emotion model to demonstrate real-time emotion recognition in the browser.

---

If you want, I can also create requirements.txt, add a short README badge header, or help containerize this with Docker.