Emotion recognition from facial images using convolutional neural networks.
Resnet, VGG-16, and Onsunet convolutional neural network architectures are used.
PyTorch, TensorFlow libraries are used to implement convolutional neural networks. Each architecture has two variants for each libary.
PySide is used to provide graphical user interface to application.
FER2013, KDEF, and CK+ datasets are used to train classifier. Each dataset is processed in both PyTorch and TensorFlow libraries.
To execute the application, it is enough to run main.py file.
python main.py