Emotion Recognition

The project entails the creation of an artificial intelligence application, designed to empower the marketing department in gauging the emotional responses of individuals engrossed in their advertisements. This innovative system is engineered to deduce the emotional states from facial expressions captured during the viewer's interaction with the ad. The application's core functionality is to classify the observed emotional states into seven distinct categories: anger, disgust, fear, joy, neutral, sadness, and surprise, employing the sophisticated capabilities of Convolutional Neural Networks (CNN) by capturing the video frames of the human faces and classifying them uing the trained model.

Running the App

To compile from the source:

• Anaconda Python 3.9 for the specific platform is required and can be found here: https://docs.anaconda.com/free/anaconda/reference/packages/py3.9_win-64/ For a windows machine. Packages of Linux and Mac can also be found on the Anaconda Python website.

After downloading Anaconda Python.

• Clone the repository: https://github.com/takavarasha-desire/realTime7

git clone https://github.com/takavarasha-desire/realTime7
cd realTime7

• Make your virtualenv

virtualenv env

• Activate the virtualenv On Windows

env/scripts/activate

On MacOS and Linux

source env/Scripts/activate

Install cv2

pip install opencv-python

Install keras

pip3 install keras

Run the app

python application.py .\haarcascade_frontalface_default.xml