Vision-Based Human Emotion Recognition Using HOG-KLT Feature

Abstract

Obviously, humans communicate and interact among each other through speech and body movement. It is also obvious that there is a close linkage between human emotion expressions and his body movements. This implies that emotion is an important aspect in the interaction and communication between people. Since the science of artificial intelligence (AI) is concerned with the automation of intelligent behavior. This paper aims to recognize the emotion of the human using histogram of orientation gradient (HOG) and Kanade–Lucas–Tomasi (KLT) HOG-KLT feature. The basic emotions used in this work are angry, joy, fear, sad and pride. The input videos are converted into gray frames. The HOG-KLT features are extracted from the sequences of frames. The emotions are recognized using support vector machine and random forest classifier. The GEMEP corpus dataset is used for this experiment.