Intelligent Facial Emotion Recognition using modified-PSO

In this paper, a facial emotion recognition technique is proposed. The proposed technique employs modified-LBP, which conduct horizontal and vertical neighbourhood pixel comparison, to generate a discriminative initial facial representation. Then, a modified-PSO, is proposed to perform feature optimization.[2] The Facial emotion recognition system consists of three steps: 1) Feature Extraction; 2) Feature Optimization; 3) Emotion Recognition. Firstly, we use modified local binary patterns (LBPs), i.e., horizontal and vertical neighbourhood comparison LBP, to extract the initial facial representation. Then, the proposed PSO algorithm is used to identify the most discriminative and significant features for differentiating distinct facial expressions. SVM(Support Vector Machine) and Multiple-SVM classifiers are used for recognizing six facial expressions.: 1) Happiness; 2) Sadness; 3) Anger; 4) Fear; 5) Surprise; 6) Disgust.