Department of Information Management, Hunan University of Finance and Economics, 410205, China

---

Abstract—In this paper, We use Adaboost to create MILBoost and propose a new MILBoost approach to automatically recognize the facial expression from video sequences by constructing the MILBoost methods. At first, we determine facial velocity information using optical flow technique, which is used to charaterize facial expression. Then visual words based on facial velocity is used to represent facial expression using Bag of Words. Final MILBoost model is used for facial expression recognition, in order to improve the recognition accuracy, the class label information was used for the learning of the MILBoost model. Experiments were performed on a facial expression dataset built by ourselves and evaluated the proposed method, the experiment results show that the average recognition accuracy is over 89.2%, which validates its effectiveness.

Index Terms—Facial expression recognition; Motion feature; Bag of Words; MILBoost

[PDF]

Cite: Shaoping Zhu, "Facial Expression Recognition Based on MILBoost," Journal of Software vol. 9, no. 9, pp. 2435-2442, 2014.