Facial Expression Recognition: Machine Learning Algorithms and Feature Extraction Techniques

Abstract

Facial expression recognition (FER) systems accurately identify facial expressions by extracting facial features. The extraction of robust facial features comes after automatic face detection in this procedure. On the FAR2013 dataset, a five-step system developed to assess the performance of machine learning algorithms. Components of the system include preprocessing, feature extraction, model training and testing, classification, and evaluation. Three machine-learning algorithms utilized in this study: logistic regression (LR), random forest (RF), and AdaBoost (ADA). The RF algorithm achieved the highest degree of precision with a 61% success rate. The purpose of the study was to evaluate the performance of machine learning algorithms on the FAR2013 dataset. The study highlights the importance of facial feature extraction in FER systems and the precision of machine learning algorithms in facial expression recognition.

Keywords

• Face expression; Machine learning, Feature extraction, Random Fore, logistic regression. PCA, GLCM