Face Emotion Classification using AMSER with Artificial Neural Networks

The Facial expression recognition phases provides a leading challenges and issues which are volatile while representing the different facial expression. The test sample complexity is not measured during the training phase of classification model. In certain parts of the faces the emotions are detected by the physiological changes that are expressed externally on face. The human cognition laws like principle of simplicity are unreliable in finding out the suitable test samples. To overcome, a new algorithm is proposed Advanced Maximally Stable Extremal Regions (AMSER) method for extracting the features. With the corresponding dataset and Artificial Neural Networks (ANN), the classification process is extracted with better facial expressions. The experimental result shows the perfectness and correct accuracy in classifying the facial expressions.