Performance comparisons of artificial neural network algorithms in facial expression recognition

This paper presents methods for identifying facial expressions. The objective of this paper is to present a combination of texture oriented method with dimensional reduction and use for training the Single-Layer Neural Network (SLN), Back Propagation Algorithm (BPA) and Cerebellar Model Articulation Controller (CMAC) for identifying facial expressions. The proposed methods are called intelligent methods that can accommodate for the variations in the facial expressions and hence prove to be better for untrained facial expressions. Conventional methods have limitations that facial expressions should follow some constraints. To achieve the expression detection accuracy, Gabor wavelet is used in different angles to extract possible textures of the facial expression. The higher dimensions of the extracted texture features are further reduced by using Fisher’s linear discriminant function for increasing the accuracy of the proposed method. Fisher’s linear discriminant function is used for transforming higher-dimensional feature vector into a two-dimensional vector for training proposed algorithms. Different facial emotions considered are angry, disgust, happy, sad, surprise and fear are used. The performance comparisons of the proposed algorithms are presented.