Facial Image Processing in Computer Vision

Background

The study in facial expression has been conducted in last century (Darwin, 1872; P. Ekman, 1973), and within the past 12 years considerable progress has been made in automatic analysis of facial expression from digital video input (J.F Cohn, 2007; J.F. Cohn & Kanade, 2007; Fasel & Luettin, 2003; Tian, Cohn, & Kanade, 2005). In early research, Ekman et al (P. Ekman & Friesen, 1976) reported a new method of describing facial movement based on an anatomical analysis of facial action. To capture subtlety of human emotion and paralinguistic communication, automated recognition of fine-grained changes in facial expression is needed (Tian, Kanade, & Cohn, 2001).

Early researches in automatic analysis and recognition of facial actions from input video focused on the relatively tractable problem of posed facial actions acquired under well-controlled conditions. Recent work has progressed to spontaneous facial actions, subtle facial actions, variation in illumination and merging with synthesis faces (i.e. avatar separate identity from facial behaviour) (J.F Cohn, 2007). Computer facial expression analysis systems need to analyze the facial actions regardless of context, culture, gender, and so on. The accomplishments in psychological studies, human movement analysis, face detection, face tracking, and recognition motivate the automatic facial expression analysis (Tian, et al., 2005).