Facial Expression Recognition Using Shearlet Transform and Kirsch Masking

In this paper, a new feature extractor is proposed that is combined of Discrete Shearlet Transform (DST) system and Kirsch Compass Kernel (KCK) with Local Binary Pattern (LBP) for extracting features from images. Discrete Shearlet Transform gives multiresolute images of an image by highlighting the major edges along with detail features. On the other hand, Kirsch Compass Kernel (KCK) provides 8 different directional images with prominent edges which sharpen the edges in each direction. These characteristics of DST and KCK help LBP to extract the significant features of an image. The extracted features by LBP help to classify the images by using Support Vector Machine (SVM).