Face Detection and Recognition of the Seven Emotions via Facial Expression: Integration of Machine Learning Algorithm into the NAO Robot

This study revolves around NAO, a programmable and interactive robot, to do Emotion Recognition via Facial Expression, and to give an appropriate response for the identified emotion whilst accumulating images to further enhance the accuracy. To develop this cohesive system, the authors utilized Comp...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2020 5th International Conference on Control and Robotics Engineering (ICCRE) S. 25 - 29
Hauptverfasser: Rosula Reyes, S.J., Depano, Keanu M., Velasco, Aaron Matthew A., Kwong, John Chris T., Oppus, Carlos M.
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.04.2020
Schlagworte:
Choregraphe Suite
Emotion recognition
Emotion Recognition via Facial Expression
Face detection
Feature extraction
Histogram of Oriented Gradien
Human-Robot Interaction
Local Binary Pattern
Machine learning algorithms
NAO
NAOqi
Robot kinematics
Support Vector Machine
Support vector machines
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This study revolves around NAO, a programmable and interactive robot, to do Emotion Recognition via Facial Expression, and to give an appropriate response for the identified emotion whilst accumulating images to further enhance the accuracy. To develop this cohesive system, the authors utilized Computer Vision and Haar-Cascade Classifier along with the feature descriptors, Histogram of Oriented Gradient and Local Binary Pattern to train the Support Vector Machine Learning Algorithm. Gathering and increasing data with Asian/Filipino participants whilst being added to the CK+ database in different stages of retraining, optimization, and undertaking cross-fold validation yielded improved results with the highest average accuracy across the seven emotions of 87.14%. Suggesting that adding images with a specific ethnicity yields a higher accuracy model for a more diverse testing dataset and in the wild image classifications. With the machine learning model capable of accurately recognizing the emotion and reinforcing it with the accumulated local images, NAO equipped with emotion recognition capabilities can better aid and support the individuals in need.
DOI:10.1109/ICCRE49379.2020.9096267