General Information

Bachelor Thesis

• Title: Automatic Facial Emotion Recognition through Information Fusion
• Author: Pablo Tanner (mail@pablotanner.com, feel free to message me for any questions)
• Supervisor: Prof. Dr. Manuel Günther

Dataset

AffectNet

Libraries and Frameworks

• OpenFace: For facial feature extraction
• DeepFace: For facial recognition embedding extraction
• cuML: For GPU-accelerated machine learning models
• scikit-learn: For any other machine learning models not supported by cuML
• PyTorch: For deep learning models (MLP and SequentialNN)

Important Notes

• This repository contains all the files and code used for the experiments, even if not included/utilized in the final thesis.
• Code that wasn't used should have a comment indicating so (e.g. "This was not used in final version")
• The code utilized for the visualizations can be found in /src/visualizations
• I did my best to add comments indicating in what Section/for what Experiments the different code files were used