The System detects the facial emotion among the seven different types of emotions using the convolutional neural network (CNN) Architecture at the accuracy rate of 60 % and a loss of less than 0.20.
The seven types of expressions are :
Angry
sad
fear
neutral
surprise
disgust
Happy
The model is trained and tested over fer_2013 image dataset available on kaggle, with the total epoche count of 50.
The model is connected with the web0cam using OpenCV.
To clone the repository in your system use :
git clone https://github.com/RajKhanke/facial-expression-detection-using-Machine-Learning.git
dataset link : https://www.kaggle.com/datasets/msambare/fer2013