Research On Independent Emotion Recognition From Facial Expression

Abstract

Emotions are essential to human communication. It is also helpful to understand the person's attitude and behavior. This study suggests a deeper learning approach to perceiving emotions from the face. We have selected two methods of convolution neural network (CNN) and Artificial neural network (ANN). We will work in all cases depending on the autonomy / dependence. We aim here to see the independent feelings of the subject well. Recent FER focuses on independent emotional recognition and has used the JAFFE dataset (Japanese women's expression) as a benchmark for benchmarking. We have chosen to use voting classifier as our benchmark algorithm and are trying to get better results than in the past. We say the best because it is not the limited data used by the previous paper while it is a combination of learning integration that works with multiple variables and predicts the outcome in terms of high class opportunities. We use the confusion matrix to test the performance of our model..