1. 1: Introduction
Face plays important role in social communication. This is a 'windows' to real human identity, reactions and ideas. the psychological research shown that nonverbal part is the most enlightening route in interpersonal communication. Verbal part offers about 7% of the note, vocal - 34% and cosmetic expression about 55%.
Due compared to that, face is a style of analysis in many areas of science such as mindset, behavioral science, medication and finally computer science. In the field of computer technology much effort is put to find the means of automation the procedure of face recognition and segmentation. Several methods dealing with the condition of cosmetic feature removal have been proposed. The key problem is to provide ideal face representation, which leftovers strong with respect to diversity of cosmetic appearances.
The method of face recognition takes on an important role in people's life ranging from commercial to police applications, such as real time security, biometric personal id and information security. It really is one of the most challenging topics in the software of computer eye-sight and cognitive research. Over past years, intensive research on face reputation has been conducted by many psychophysicists, neuroscientists and engineers. Generally views, the definition of face recognition can be designed as follows Different faces in a static image can be identified using a data source of stored faces. Available guarantee information like cosmetic expression may enhance the recognition rate. Generally speaking, if the face images are sufficiently provided, the grade of face popularity will be mainly related to feature extraction and popularity modeling.
Facial emotion recognition in uncontrolled surroundings is an extremely challenging task credited to large intra-class versions caused by factors such as illumination and create changes, occlusion, and head movement. The accuracy of a cosmetic emotion acknowledgement system generally depends upon two critical factors: (i) removal of cosmetic features that are robust under intra-class versions (e. g. pose changes), but are distinctive for various feelings, and (ii) design of a classifier that is with the capacity of distinguishing different facial emotions predicated on noisy and imperfect data (e. g. , brightness changes and occlusion).
For identification modeling, plenty of researchers usually evaluate the performance of model by popularity rate rather than computational cost. Recently, Wright and Mare ported their work called the sparse representation structured classification (SRC). To be more specific, it can stand for the testing image sparsely using training examples via norm minimization that can be solved by controlling the least reconstructed error and the sparse coefficients. The acceptance rate of SRC is much higher than that of traditional algorithms such as Nearest Neighbor, Nearest Subspace and Linear Support Vector Machine (SVM). However, there are three downsides behind the SRC. First, SRC is based on the all natural features, which cannot exactly catch the incomplete deformation of the face images. Second, regularized SRC usually works slowly for high dimensional face images.
Third in the presence of occluded face images, Wright et al introduce an occlusion dictionary to sparsely code the occluded components in face images. However, the computational cost of SRC increase considerably because of large number of elements in the occlusion dictionary. Therefore, the computational cost of SRC limits it s program in real time area, which progressively more attracts researcher's focus on solve this issue.
1. 2: Psychological Background
In 1978, Ekman et al.  presented the system for measuring cosmetic expressions called FACS - Cosmetic Action Coding System. FACS originated by evaluation of the relationships between muscle(s) contraction and changes in the facial skin appearance triggered by them. Contractions of muscles in charge of the same action are proclaimed as an Action Device (AU). The duty of expression evaluation with use of FACS is based on decomposing observed manifestation into the set of Action Units. You will discover 46 AUs that represent changes in cosmetic appearance and 12 AUs connected with eyeball gaze direction and mind orientation. Action Systems are highly descriptive in conditions of facial motions, however, they don't provide any information about the concept they signify. AUs are tagged with the description of the action (Fig. 1).
Fig. 1: Examples of Action Units
Facial expression referred to by Action Systems can be then examined on the semantic level and discover the meaning of particul