Open AccessEfficient CNN Approach for Facial Expression Recognition
Author(s) -
Gheyath Mustafa Zebari,
Dilovan Asaad Zebari,
Diyar Qader Zeebaree,
Habibollah Haron,
Adnan Mohsin Abdulazeez,
Kamil Yurtkan
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2129/1/012083
Subject(s) - convolutional neural network , facial expression recognition , computer science , artificial intelligence , facial expression , field (mathematics) , deep learning , pattern recognition (psychology) , artificial neural network , expression (computer science) , convolution (computer science) , machine learning , emotion recognition , facial recognition system , speech recognition , mathematics , pure mathematics , programming language
In the last decade, the Facial Expression Recognition field has been studied widely and become the base for many researchers, and still challenging in computer vision. Machine learning technique used in facial expression recognition facing many problems, since human emotions expressed differently from one to another. Nevertheless, Deep learning that represents a novel area of research within machine learning technology has the ability for classifying people’s faces into different emotion classes by using a Deep Neural Network (DNN). The Convolution Neural Network (CNN) method has been used widely and proved as very efficient in the facial expression recognition field. In this study, a CNN technique for facial expression recognition has been presented. The performance of this study has been evaluated using the fer2013 dataset, the total number of images has been used. The accuracy of each epoch has been tested which is trained on 29068 samples, validate on 3589 samples. The overall accuracy of 69.85% has been obtained for the proposed method.