Welcome.

Empathy is a small computer vision and deep learning project for the Digital Image Processing course at FCSE. Empathy detects faces (with Haar Cascades) in video and images and describes their emotion using a trained/to-be-trained CNN.

Please note that the app comes with a preloaded and trained model.

Required packages:

NumPy (numpy),

PyTorch (torch),

Keras preprocessing library (keras-preprocessing),

OpenCV (opencv-python) for Python,

tkinter,

Pillow (pillow) and

Augmentor

If you would like to train the model from scratch, please download the FER2013 dataset from https://www.kaggle.com/datasets/deadskull7/fer2013 and put it in the datasets (create it if it doesn't exist) directory, then do the following:

1. Run the preprocessing/dataset_formatting.py script (if you are using a different dataset than FER2013, make sure it matches FER2013's column and emotion number format)
2. Run the preprocessing/augumentation.py stript
3. Run the preprocessing/training.py script

Authors: Jana Angjelkoska, David Dukoski