Page 78 - Read Online

Bah et al. Intell Robot 2022;2(1):72­88 Intelligence & Robotics
DOI: 10.20517/ir.2021.16

Research Article Open Access

Facial expression recognition using adapted residual
based deep neural network

1
Ibrahima Bah , Yu Xue 1,2
1 School of Computer and Software, Nanjing University of Information Science and Technology, Nanjing 210044, Jiangsu, China.
2 Jiangsu Key Laboratory of Data Science and Smart Software, Jingling Institute of Technology, Nanjing 211169, Jiangsu, China.

Correspondence to: Dr. Ibrahima Bah, School of Computer and Software, Nanjing University of Information Science and Tech-
nology, No. 219, Ningliu Road, Pukou District, Nanjing 211169, Jiangsu, China. E-mail: 20205220003@nuist.edu.cn; Prof. Yu Xue,
School of Computer and Software, Nanjing University of Information Science and Technology, No. 219, Ningliu Road, Pukou District,
Nanjing 211169, Jiangsu, China. E-mail: xueyu@nuist.edu.cn;
How to cite this article: Bah I, Xue Y. Facial expression recognition using adapted residual based deep neural network. Intell Robot
2022;2(1):xx. http://dx.doi.org/10.20517/ir.2021.16
Received: 6 Dec 2021 First Decision: 21 Feb 2022 Revised: 24 Feb 2022 Accepted: 3 Mar 2022 Published: 22 Mar 2022

Academic Editor: Simon X. Yang Copy Editor: Xi-Jun Chen Production Editor: Xi-Jun Chen

Abstract
Emotion on our face can determine our feelings, mental state and can directly impact our decisions. Humans are
subjected to undergo an emotional change in relation to their living environment and or at a present circumstance.
These emotions can be anger, disgust, fear, sadness, happiness, surprise or neutral. Due to the intricacy and nuance
of facial expressions and their relationship to emotions, accurate facial expression identification remains a difficult
undertaking. As a result, we provide an end-to-end system that uses residual blocks to identify emotions and improve
accuracy in this research field. After receiving a facial image, the framework returns its emotional state. The accuracy
obtained on the test set of FERGIT dataset (an extension of the FER2013 dataset with 49300 images) was 75%. This
proves the efficiency of the model in classifying facial emotions as this database poses a bunch of challenges such
as imbalanced data, intraclass variance, and occlusion. To ensure the performance of our model, we also tested it on
the CK+ database and its output accuracy was 97% on the test set.

Keywords: Facial expression recognition, emotion detection, convolutional neural network, deep residual network

© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0
International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, shar­
ing, adaptation, distribution and reproduction in any medium or format, for any purpose, even commercially, as long as you
give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate
if changes were made.

www.intellrobot.com