An Affective Computing course project at the University of Oulu, built on the DISFA dataset.
I trained and evaluated a ResNet model for facial expression recognition, focused on detecting facial action units: the small, codified facial movements that expression analysis builds on.