A Novel Approach to Face Pattern Analysis

Recognizing facial expressions is a major challenge and will be required in the latest fields of research such as industrial Internet Things. Currently, available methods are useful for detecting singular images, but they very hard to extract. The main aim face detection capture an image real-time search dataset. So, by using this biometric feature, one can recognize verify person’s their features. Many researchers have used Principal Component Analysis (PCA), Support Vector Machine (SVM), combination PCA SVM, with Artificial Neural Network, even traditional PCA-SVM improve recognition. better than PCA-ANN has limitation small As far classification generalization concerned, SVM requires fewer parameters generates less errors ANN. In paper, we propose new framework, called FRS-DCT-SVM, that uses GA-RBF optimization discrete cosine transform (DCT) extract FRS-DCT-SVM gives results terms clustering time. average accuracy received 98.346, which SVM-DCT (86.668 96.098, respectively). addition, comparison made based on training, testing, times.