Image Emotion Classification using deep learning

Abstract

Behavioral poses and facial expressions and recognizing them is an interesting field of research. Humans have various emotions with multiple intentions so tremendous research has undergone to understand and analyze the emotion. In this paper, we have presented a method for image emotion detection under a study of facial expression analysis. The neural network solutions and the image processing technique are used to recognize and classify the face image expression: Happy, Neutral, Cry, and Angry. Color images of the face are taken as the input. The algorithm that we use here to implement the face image emotion deduction is Convolution Neural Network. Then we have data set images in the form of pixels with a list of trained datasets of emotions. After image-processing with multiple features extract a set of feature points from the images of the face, and training and testing data sets are split by using these features. By using various classifiers, we train the model and find the accuracy of the model. Finally, we use algorithms to predict the image emotion.