Facial Expression Recognition by Considering Nonuniform Local Binary Patterns

Recognizing a face with an expression has paying attention due to its well-known applications in a broad range of fields like data-driven animation, human–machine interaction, robotics, and driver fatigue detection. People can vary significantly their facial expression; hence, facial expression recognition is not an easy problem. This paper presents a significant contribution for facial expression recognition by deriving a new set of stable transitions of local binary pattern by selecting the significant nonuniform local binary patterns. The proposed patterns are stable, because of the transitions from two or more consecutive ones to two or more consecutive zeros. For better recognition rate, the new set of patterns are combined with uniform patterns of local binary pattern. A distance function is used on proposed texture features for effective facial expression recognition. Preprocessing method is also used to get rid of the effects of illumination changes in facial expression by preserving the significant appearance details that are needed for facial expression recognition. The investigational analysis was done on the popular JAFFE facial expression database and has shown good performance.