Emotion Recognition

Emotion Recognition Model using the fer2013 dataset, built with Keras and OpenCV, with a 70% accuracy

• Data processing files borrowed from here
• MTCNN package borrowed from here

To run the program, open command line and type:

python3 emotion_color_demo.py

To run the program with the MTCNN face detection neural network instead of OpenCV's Haar Feature Classifer, run:

python3 mtcnn_demo.py

To train your own model, download the fer2013 package from here Create a "datasets" folder and unzip the file under it:

tar -xzf fer2013.tar

Run the training program:

python3 train_emotion_classifer.py