EmoDetect

Real-Time Emotion Detection App

This project is a Real-Time Emotion Detection application built with Python, OpenCV, and Keras. It captures video from a webcam, detects faces, and classifies emotions in real-time, specifically tracking “happy” expressions for a duration of approximately 3 seconds. When a “happy” face is detected continuously for this threshold, a message is displayed on the video feed.

Features

• Face Detection: Uses OpenCV’s Haar Cascade Classifier to detect faces in real-time.
• Emotion Classification: Classifies emotions into predefined categories: “Angry,” “Disgust,” “Fear,” “Happy,” “Neutral,” “Sad,” and “Surprise” using a pre-trained model.
• Consecutive Happy Detection: Counts consecutive frames with a “happy” expression. If “happy” is detected for ~3 seconds, a message is displayed.
• Real-Time Video Display: Continuously displays the video feed, showing emotions and tracking the duration of “happy” detections.
• Simple Exit Control: User can end the program by pressing the q key.

Requirements

To run this project, you need the following libraries:

• Python 3.x
• OpenCV
• Keras
• TensorFlow
• NumPy

Install dependencies with:

pip install opencv-python keras tensorflow numpy

File Structure

• emotion_detection.py: Main script for running the emotion detection application.
• haarcascade_frontalface_default.xml: Haar Cascade Classifier file for face detection.
• model.h5: Pre-trained model for emotion classification.

Usage

Clone the Repository

git clone https://github.com/haffarsadok/EmoDetect.git
cd emotion-detection-app

Interact with the Application

• The app will open a real-time video feed from your webcam.
• It will display detected emotions for each face.
• If a “happy” face is detected for ~3 seconds, a message “Happy detected for 3 seconds!” will appear on the screen.
• Press the q key to exit the application.

Model Details

The emotion detection model is a convolutional neural network (CNN) trained on facial emotion datasets. It classifies emotions into the following categories:

• Angry
• Disgust
• Fear
• Happy
• Neutral
• Sad
• Surprise

Customization

To adjust the duration for which a “happy” expression must be detected, modify the happy_duration_threshold variable in emotion_detection.py.

Overview

Contributing

We welcome contributions! Feel free to open issues or submit pull requests for improvements or additional features.

License

This project is licensed under the MIT License.

This site is open source. Improve this page.