A Comparison of the Performance of Artificial Neural Network Algorithms in Facial Expression Recognition | Chapter 1 | Current Approaches in Science and Technology Research Vol. 12

In this study, the methodologies for distinguishing facial expressions are described. The purpose of this research is to show how to train the Single-Layer Neural Network (SLN), Back Propagation Algorithm (BPA), and Cerebellar Model Articulation Controller (CMAC) for identifying facial expressions using a texture-oriented strategy combined with dimensional reduction. Because the proposed approaches can account for changes in face expressions and so perform better for untrained facial expressions, they are referred to as intelligent methods. Face expressions must follow to particular rules, which limits traditional methods. To achieve expression identification accuracy, the Gabor wavelet is used from various perspectives to extract probable textures of the face expression. To improve the accuracy of the suggested method, the higher dimensions of the collected texture characteristics are further decreased using Fisher's linear discriminant function. Fisher's linear discriminant function is used to convert a higher-dimensional feature vector into a two-dimensional vector for training the proposed algorithms. Some of the facial emotions employed are anger, disgust, happiness, sadness, surprise, and fear. The performance of the proposed algorithms is compared.

Author (S) Details

Amira Elsir Tayfour
King Khalid University, Saudi Arabia.

Altahir Mohammed
Sudan University of Science & Technologies, Sudan.

Moawia Elfaki Eldow
University of Khartoum, Sudan.

View Book :- https://stm.bookpi.org/CASTR-V12/article/view/2593