Gaze Recognition Based on Correlation of Facial Components

In this paper, we propose a fast and accurate gaze recognition method by using the correlation of a small amount of computation facial components. The proposed method detects the face, eyes, nose and mouth. In the detected face region, using the Haar-like features, the center point of the detected area, the distance between each of the components, and each of the components was extracted through the angle of the three corners of the triangle in the components of the detected face. The extracted features in the learning set were used for recognition experiments with the Random Forest algorithm, achieving a very high recognition rate of 99.46% for the recorded results. In addition, the perform rate was demonstrated to be 23.6fps.