Facial expression recognition using local binary pattern and modified hidden Markov model Online publication date: Fri, 11-Sep-2020
by Mayur Rahul; Narendra Kohli; Rashi Agarwal
International Journal of Advanced Intelligence Paradigms (IJAIP), Vol. 17, No. 3/4, 2020
Abstract: Facial expression is the non-verbal communication used to send and receive their inner emotions state. It plays a key role in person's behaviour and communication. Facial expression recognition can be applied in various areas like human-computer interaction, surveillance systems, medicines, home security and credit card verification. An efficient and robust face descriptor is very important in FER systems. This paper is able to analyse the local binary pattern (LBP) to represent the facial features in face images. LBP can be calculated in all eight directions of each pixel to obtain the binary coded number. Each expression is represented using binary code. We incorporated this LBP with our new modified HMM, which acts as a classification method. New modified HMM classifier is made up of two layers: upper layer and bottom layer. Bottom layer consists of atomic expressions and upper layer consists of combinations of atomic expressions. There are seven classes of known expression, i.e., anger, disgust, fear, joy, sadness, surprise, neutral are recognised with this approach. We have also tested our new framework using confusion matrix, ROC curve, recognition performance, processing time, error rates and found the overall accuracy of 85%.
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Advanced Intelligence Paradigms (IJAIP):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email subs@inderscience.com