RIQ-FSP Model for Face Recognition

Face recognition has become a very important research area of image processing, which is used in a variety of fields and disciplines. The paper presents a new model for face recognition. The model uses a hybrid color space (HCS) RIQ, which has been constructed by combining color components of RGB color space and YIQ color space. For feature reduction the model uses Fisher Linear Discriminant (FLD) method along with Face Symmetry Property (FSP). When FLD was combined with FSP, the model was named as FLD-FSP. The performance of the FLD-FSP model was estimated using face images of 64×64 resolution. The face images were extracted by employing Viola John face detector. The results show that the system can perform considerably well in real world controlled environments.