Emotion Recognition by analysing statistical features of cropped eyebrow portion as a Region of Interest of facial expression image

摘要

In this paper, an approach of emotion recognition is carried out by analysing statistical features. In the proposed work firstly on JAFFE standard database ROI considered. Since eyebrow, mouth, nose, left eye, and right eye these portions of facial image plays vital role in finding human emotions. The cropped eyebrow images of different emotions of facial images are considered for calculating statistical feature. In this proposed work unique statistical feathers have been calculated to identify specific emotion of human being. The statistical features like mean, median, mode, standard deviation median Abs. Dev and mean Abs. Dev Calculated for each eyebrow image. By analysing these statistical features human emotions like happy, sad, angry, fear, disgust, surprise and neutral recognizes very accurately.