Histogram of Orientation Gradient Feature-Based Facial Expression Classification Using Bagging with Extreme Learning Machine

摘要

Facial expression analysis is widely used in behavior interpretation of emotion, cognitive science, and social interaction. This paper presents a method for facial expression classification based on the Histogram of Orientation Gradient (HOG) features and Extreme Learning Machine (ELM) alone with bagging algorithm. After detecting the face from the image, the HOG features are detected from the face image by dividing it into number of overlapping rectangular blocks. ELM is used for training the dataset. In order to improve the classification results, the bagging method is used, which significantly improves the classification performance even with fewer base classifiers. The results of our proposed method are better than existing methods in the literature.