Low-Resolution Face Recognition Using Unimodal Data Fusion

The objective of low-resolution face recognition is to identify faces in an uncontrolled situations like from small size or poor quality images with varying pose, illumination, expression, etc. Most existing approaches use features of just one type. In this work, we propose a robust low face recognition technique based on unimodal features fusion, which is more discriminative than using only one feature modality. Features of each facial image are extracted using three steps: i) both Gabor filters and Histogram of Oriented Gradients (HOG) descriptor are calculated. ii) the size of these features is reduced using the Linear Discriminant Analysis (LDA) method in order to remove redundant information. iii) the reduced features are combined using Discriminant Correlation Analysis (DCA) method. To achieve the recognition task, a Support Vectors Machine Classifier, is used. Performance of the proposed method will be measured using the AR database