Human Expression Recognition using Facial Shape Based Fourier Descriptors Fusion

摘要: Dynamic facial expression recognition has many useful applications in social networks, multimedia content analysis, security systems and others. This challenging process must be done under recurrent problems of image illumination low resolution which changes at partial occlusions. paper aims to produce a new method based on the muscles. The geometric features are used specify regions i.e., mouth, eyes, nose. generic Fourier shape descriptor conjunction with elliptic is as an attribute represent different emotions frequency spectrum features. Afterwards multi-class support vector machine applied for classification seven human expression. statistical analysis showed our approach obtained overall competent using 5-fold cross validation high accuracy well-known dataset.