Title: Facial emotion detection using convolutional neural network algorithm

Authors: G.R. Karpagam; B.S. Balasarath; Jeffrey Y. Nicholas; R. Lokesh; Shibi S. Rahul; Souradrita Sarkar

Addresses: Computer Science and Engineering Department, PSG College of Technology, Coimbatore – 641004, Tamil Nadu, India ' Computer Science and Engineering Department, PSG College of Technology, Coimbatore – 641004, Tamil Nadu, India ' Computer Science and Engineering Department, PSG College of Technology, Coimbatore – 641004, Tamil Nadu, India ' Computer Science and Engineering Department, PSG College of Technology, Coimbatore – 641004, Tamil Nadu, India ' Computer Science and Engineering Department, PSG College of Technology, Coimbatore – 641004, Tamil Nadu, India ' Computer Science and Engineering Department, PSG College of Technology, Coimbatore – 641004, Tamil Nadu, India

Abstract: Emotions form a part and parcel of our human life. As emotions are a direct reflection of our inner self, it is possible to understand the state of mind of people. Some of the common human emotions are happy, sad, anger, disgust, fear and neutral. Several types of research have been done in the field of facial expression detection out of which the method of using a convolutional neural network proved to be the most reliable and accurate. This paper makes use of CNN architectures like Inception Resnet V2 and Xception to predict the emotion of psychiatric patients from the input video file. The results of the prediction will then be published in the form of graphs in a pdf file.

Keywords: convolutional neural network; CNN; psychiatric patients; facial emotions; deep learning; Inception Resnet v2; Xception; emotion detection; comparison of architectures; Caffe model.

DOI: 10.1504/IJAIS.2022.10049037

International Journal of Adaptive and Innovative Systems, 2022 Vol.3 No.2, pp.119 - 134

Received: 28 Jun 2021
Accepted: 10 Dec 2021

Published online: 25 Jul 2022 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article