Emotional Facial Expressions using Cat Swarm Optimization

Abstract

Many research and professional applications rely on recognition of facial expression. Under this study, the group presented a novel technique for identifying emotional expressions. First, designers extracted transformation function from face expressions using the wavelet transforms. Second, to reduce the characteristics, feature extraction has been used. Third, the classifier was indeed a singular mathematical computation. Eventually, and perhaps most notably, designers used Cat Swarm Optimization (CSO) to practice the classifier's weight values. The cat swarm optimization technique obtained an average precision of 90%, 0.76 percent after ten-fold stratified classification technique. It outperformed the neural network, particle swarm, and genetic algorithms with moment acceleration coefficients. Furthermore, our facial expression processing device outperformed two state-of-the-art methods.