LBP-HOG-Statistical-Wavelet Transform Feature Based MCA Classifier

Sneha V. V

Abstract

Face recognition systems use computer algorithms choose specific, recognizable features on a person’s face. With a mathematical representation, the information is compared to data on other faces acquired in a face recognition database. The distance between the eyes or the shape of the chin are two examples of these characteristics. The job of matching several facial modes, such as visible and near infrared images, is known as heterogeneous face recognition. The main problems with HFR are the insufficient training samples and the severe lack of modality compatibility. Since the same person’s face images from several image modalities are linked to the same face object, there need to be a mutual component that reflects the intrinsic features of the face that are independent of the image modalities. To infer the mutual components for reliable heterogeneous face recognition, a Mutual Component Analysis (MCA) is proposed in this study and also tried other feature extraction methods to compare the performance. This performance is shown by using NB classifier

Abstract

Keywords