/FacialEmotionDetection

Explore our Flask-based web app for real-time facial emotion recognition using advanced CNN models. Dive into dynamic emotion analysis and visualization, ideal for developers and researchers.

Primary LanguageHTMLMIT LicenseMIT

Real-time Facial Emotion Recognition Web Application

This repository hosts a cutting-edge web application designed for real-time facial emotion recognition, blending Flask, Python, HTML, CSS, and advanced machine learning models. It's distinguished by its use of three pivotal models: Convolutional Neural Network (CNN), Chehra model, and Deep Neural Network (DNN), each meticulously trained for high accuracy in detecting diverse emotional states. A hallmark of this project is its interactive visualization, which dynamically showcases emotion detection results and graphically represents them for insightful analysis.

Project Overview

Understanding and analyzing human emotions in real time can significantly impact various domains such as mental health assessment, user experience enhancement, and educational tools. Our web application offers a robust and user-friendly platform for real-time analysis and categorization of human emotions through facial expressions, leveraging cutting-edge AI and human-computer interaction advancements.

Key Features

  • Real-time Emotion Detection: Detect and analyze facial emotions in live video streams using your web camera.
  • Interactive Visualization: Dynamic display of detected emotions on the web page, along with a bar chart graph for an in-depth analysis over time.
  • Model Flexibility: Switch seamlessly between CNN, Chehra, and DNN models, catering to various needs for accuracy and computational efficiency.
  • User-Friendly Interface: An intuitive and navigable front-end ensures a smooth user experience.
  • Comprehensive Reporting: Generate detailed reports of detected emotions for further analysis.

Motivation

The project aims to bridge the gap between human emotions and technology, providing insights into emotional dynamics and enhancing interactions in digital spaces. It stands as a testament to the potential applications of machine learning in real-time emotion detection and interactive visualization.

Here's the revised Installation and Setup section with an added step for setting up a virtual environment:

Installation and Setup

  1. Clone the repository:

    git clone https://github.com/gmMustafa/facial_emotion_detection/

  2. Download the architecture and model zip file from the following URL, and extract it into the root folder of the cloned repository:

    https://tinyurl.com/439nt2th

  3. Set up a Python virtual environment: Navigate to your project directory and create a virtual environment. Activate it before proceeding to the next steps.

    • Create the virtual environment: 
      python -m venv venv
    • Activate the virtual environment:
      • On Windows, run: 
        venv\Scripts\activate
      • On macOS/Linux, run: 
        source venv/bin/activate

  4. Install dependencies: With the virtual environment activated, install the required packages:

    pip install -r requirements.txt

  5. Launch the application: Ensure the virtual environment is still activated, then start the Flask application:

    python app.py

  6. Access the application by navigating to http://127.0.0.1:5000/ in your web browser.

Remember to deactivate your virtual environment when you're finished working on the project by running deactivate in your terminal.

Usage Guide

  • Grant the application permission to use your web camera.
  • Select the desired emotion detection model from the dropdown menu.
  • Experience real-time emotion detection and visualization.
  • Utilize the "Generate Report" feature for a session summary of detected emotions.

Technologies Used

  • Flask: Serves as the backbone for server-side operations.
  • HTML/CSS/JavaScript: Crafts a responsive and interactive UI.
  • TensorFlow/Keras: For model training and inference.
  • OpenCV: For video stream processing and face detection.
  • Dlib: Facial landmark detection in the Chehra model.

Contributors

A huge thank you to everyone who has contributed to this project! Your contributions help make this project better.

Screenshots

UI Overview

UI Overview

UI Overview - II

Contributing

We welcome contributions to improve functionality, model performance, or user experience. Please fork the repository and submit pull requests for review.

License

This project is licensed under the MIT License - see LICENSE for details.

Acknowledgments

Our heartfelt thanks go to the facial emotion recognition community for their research and datasets that have significantly contributed to this project's success.

Future Directions

Future enhancements will focus on integrating more advanced neural networks, refining user interfaces, and expanding the application's reach into various sectors, guided by ethical considerations and a commitment to innovation.

Disclaimer

This project is associated with the "Interactive Visualization Project" under the "AI Systems and Applications" pillar, part of the Master of Science in Artificial Intelligence degree at Friedrich-Alexander-Universität Erlangen-Nürnberg.

For a detailed exploration of the project's methodology, evaluation, and insights, refer to the presentation slides at https://tinyurl.com/4ppxptxp.

Friedrich-Alexander-Universität Erlangen-Nürnberg