A paired sparse representation model for robust face recognition from a single sample

Abstract

Sparse representation-based classification (SRC) has been shown to achieve a high level of accuracy in face recognition (FR). However, matching faces captured in unconstrained video against a gallery with a single reference facial still per individual typically yields low accuracy. For improved robustness to intra-class variations, SRC techniques for FR have recently been extended to incorporate variational information from an external generic set into an auxiliary dictionary. Despite their success in handling linear variations, non-linear variations (e.g., pose and expressions) between probe and reference facial images cannot be accurately reconstructed with a linear combination of images in the gallery and auxiliary dictionaries because they do not share the same type of variations. In order to account for non-linear variations due to pose, a paired sparse representation model is introduced allowing for joint use of variational information and synthetic face images. The proposed model, called synthetic plus variational model, reconstructs a probe image by jointly using (1) a variational dictionary and (2) a gallery dictionary augmented with a set of synthetic images generated over a wide diversity of pose angles. The augmented gallery dictionary is then encouraged to pair the same sparsity pattern with the variational dictionary for similar pose angles by solving a newly formulated simultaneous sparsity-based optimization problem. Experimental results obtained on Chokepoint and COX-S2V datasets, using different face representations, indicate that the proposed approach can outperform state-of-the-art SRC-based methods for still-to-video FR with a single sample per person.

Smooth non-negative sparse representation for face and handwritten recognition

“…1). Here, we design a different function to simultaneously apply smoothness and non-negativity, utilisable in (2).…”

“…Sparse representation problem is one of the most attractive and demanding topics in signal processing, image processing, computer vision and pattern classification research [1,2,3]. It is now explicitly observed that one can represent variety of signals/images/patterns with only few non-zero samples using an overcomplete matrix, the so-called dictionary.…”

Joint Sparsity and Collaboration Preserving Projections for Rotating Electrical Machinery Fault Diagnosis

“…In recent years, sparse representation has received considerable interest from experts and researchers in many fields, which approximately reconstructs the query sample with a combination of training samples from all classes [41]- [44]. So far, SR has been extensively employed in many theoretical researches and industrial applications, especially in the fields of signal processing [45], [46] and pattern recognition [47], [48]. The mathematical expression of SR is compact.…”

Single sample per person face recognition algorithm based on the robust prototype dictionary and robust variation dictionary construction

“…In many practical face recognition systems, for example, law enforcement, entrance monitoring, intelligent video surveillance, criminal identification, the gallery set is typically designed using only one single sample per person (SSPP) [7] because of the limited storage and privacy policy. What makes the situation even worse is that the probe samples are usually captured with video surveillance cameras under unconstrained conditions [8,9]. Therefore, SSPP face recognition in this unconstrained environment is very challenging task due to, first, the strong intraclass variations such as illumination [10][11][12], pose [12,13], facial expression [14][15][16] and occlusion [17,18], and second, the lack of gallery samples.…”