A Comparative Study of Pixel Values and Landmark Detection Features to Solve for Facial Emotion Recognition

Leonel do Nascimento, Tiago (2022) A Comparative Study of Pixel Values and Landmark Detection Features to Solve for Facial Emotion Recognition. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (5MB) | Preview
Preview	PDF (Configuration manual) Download (9MB) | Preview

Abstract

Understanding and recognizing various human emotions is foundational to how our society works. In recent decades, researchers have tried to train machine learning models that can replicate emotion recognition by utilizing Facial Emotion detection. This study aims to build different machine learning models that can effectively classify human emotions using Facial Emotion Recognition (FER) and produce a comparison between pixel values and landmark detection as features to achieve the classification. Using both posed and spontaneous datasets, different models such as Random Forest, Extra Trees classifier, Support Vector Machine and a Convolutional Neural Network are developed to achieve high accuracy. An ensemble machine learning model using hard voting created utilizing Random Forest, Extra Trees Classifier and SVM with grid search achieved 89% accuracy using pixel values as features.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science B Philosophy. Psychology. Religion > Psychology > Emotions Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Tamara Malone
Date Deposited:	22 Feb 2023 16:02
Last Modified:	02 Mar 2023 09:34
URI:	https://norma.ncirl.ie/id/eprint/6215

Actions (login required)

View Item