Facial expression classification using Cross Diagonal Neighborhood Pattern

Abstract

Facial Expression Recognition has significant applications in the field of Affective Computing. FER has its significant contribution in the fields like human computer interaction, neurology, psychiatry, image processing, computer vision, affective computing, and information security. This work gives unique and robust FER System by extracting unique and robust face features. This work proposes Cross Diagonal Neighborhood Patterns (CDNP) for unique feature extraction. The CDNP features are further processed by Gray Level Co-occurrence Matrix (GLCM). The derived CDNP-GLCM features are sent to Convolutional Neural Network (CNN) to train various expressions.