This project has been implemented in two methods:
- The first method is using a pre-trained model from the deepFace module.
- The second method involves creating and training a model using the FER-2013 dataset with 7 emotion types, which is available on Kaggle.
Both methods have the capability to detect 7 types of emotions, and the mentioned models are as follows:
- Angry
- Disgusted
- Fearful
- Happy
- Sad
- Surprised
- Neutral
Some results:
With comparing the two trained models using the same input, it can be concluded that the deepFace model has lower accuracy compared to the our trained model with the dataset.
