Facial Expression Detection Based on Local Binary Pattern and Back Propagation Neural Network
Mousa Kadhim Wali1, Md. Hussein Baqir2, Majid S. Naghmash3

1Mohammed Hussein Baqir, Master’s Degrees, Department of Electronic and Communications Engineering,  Technology University, Baghdad.
2Mousa Kadhim Wali, B.SC. Degree, Department of Electrical Engineering, Baghdad University, Iraq.
3Majid S. Naghmash, MSC. Degree, Department of Satellite Communication Engineering, Technology University, Baghdad Iraq.
Manuscript received on 11 March 2014 | Revised Manuscript received on 20 March 2014 | Manuscript Published on 30 March 2014 | PP: 72-77 | Volume-3 Issue-10, March 2014 | Retrieval Number: J15480331014/14©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Facial expression has always been used to show human’s feeling. There are many types of human’s emotion. Nevertheless, there are six primary emotions are shown in the similar way by people throughout the world regardless of culture, which are sadness, anger, happiness, fear, disgust and surprise. Since facial expressions are universal, therefore utilizing them to create several applications is possible. Face detection is crucial process where it would directly affect the emotion detection accuracy. So, this work utilizes Open CV implementation of Viola-Jones face detection library to detect faces automatically using Japanese Female Facial Expression database. Image processing and emotional classification was done in Matlab since it has excellent support tools for image processing and neural network training. Four Features namely; contrast, correlation, energy, and homogeneity were extracted based on Gray-Level Co-occurrence Matrix method after preprocessing by histogram and adaptive filter. Back propagation neural network been used in this research which yield of 87.5 % detection accuracy. A Graphical User Interface (GUI) was developed using Graphical User Interface Development Environment in Matlab.
Keywords: Facial, BPNN, LBP, GLCM.

Scope of the Article: Network Based Applications