A study on face recognition under facial expression variation and occlusion

Abstract

Two well-known problems are recognizing faces in the presence of facial expression variation and in the presence of occlusion. Humans depict their feelings through facial expressions, and this is an effective way of nonverbal communica- tion. Facial expressions are dynamic, and recognizing faces under varying facial expressions thus are the challenging task. The ability to recognize human affective state through an intelligent machine will empower to interpret, understand, and respond to human emotions, moods, and possibly intentions which is similar with one person to another. On the other hand, occlusion in an image refers to obstructions in the view of an object. Face recognition systems in real-world applications need to manage an extensive variety of obstructions, and faces can occluded by facial accessories (e.g., sunglasses, scarf, cap, cloak), objects before the face (e.g., hand, food, cellular telephone), extreme illumination (e.g., shadow), self-occlusion (e.g., non-frontal pose), or poor picture quality (e.g., blurring). In this paper, we perform a comparative study on various state of the art techniques to recognize faces under varying facial expressions and in the presence of occlusion.