Machine Learning Supported Emotion-Based Image Retrieval

Abstract

This paper proposed an emotional predictive approach for classifying emotional concepts by extracting the features of the picture vector using a deep neural convolution network. In the automated emotion prediction system where emotion is subjective, the efficiency of the prediction system can be enhanced by combining the vector features of the image instead of using a single feature such as color, texture, shape, etc. Thus, dominant color descriptors, homogeneous edge histogram descriptors, and region/contour-based descriptors are used to extract features such as color, texture, and form by three types of color factors such as activity, weight, and heat.