Assignment for Machine Learning on Multimodal Datasets.
The goal of this project is to come up with a fast and efficient solution for real-time emotion recognition on videos. It makes use of a pretrained classification model (e.g. see reference below) for the prediction of expressions. Additionally, pre-trained models from OpenCV and dlib are used, for face and facial landmark detection. This repository contains a notebook with an execution example (both for batch data, as well as a single demo). A full report on PDF is also available (in Greek).
FER2013 Dataset: https://www.kaggle.com/msambare/fer2013
CMU - MOSEI Dataset: http://multicomp.cs.cmu.edu/resources/cmu-mosei-dataset/
Code for CNN Classification on Images: https://github.com/Magda896/Project_Deep_Learning