HUMAN EMOTION DECTECTION USING HAAR FEATURES AND CONVOLUTION NEURAL NETWORKS

Abstract

Automatic Facial Expression Recognition (FER) is a part of human-machine interaction systems, which is widely researched both domestically and internationally. Current FER methods can be classified into two main groups: pure machine learning-based approaches without neural network application, and deep learning-based approaches employing convolutional neural networks (CNNs). In this study, we focus on investigating face detection using the Viola-Jones method, while implementing a convolutional neural network system to classify facial expressions, as well as combining traditional machine learning methods with modern CNNs. We conduct a comparison of results with other studies conducted over the past 5 years using the widely-used FER2013 dataset.