Robust Face Recognition via Accurate Face Alignment and Sparse Representation

Due to its potential applications, face recognition has been receiving more and more research attention recently. In this paper, we present a robust real-time facial recognition system. The system comprises three functional components, which are face detection, eye alignment and face recognition, respectively. Within the context of computer vision, there are lots of candidate algorithms to accomplish the above tasks. Having compared the performance of a few state-of-the-art candidates, robust and efficient algorithms are implemented. As for face detection, we have proposed a new approach termed Boosted Greedy Sparse Linear Discriminant Analysis (BGSLDA) that produces better performances than most reported face detectors. Since face misalignment significantly deteriorates the recognition accuracy, we advocate a new cascade framework including two different methods for eye detection and face alignment. We have adopted a recent algorithm termed Sparse Representation-based Classification (SRC) for the face recognition component. Experiments demonstrate that the whole system is highly qualified for efficiency as well as accuracy.

I. Introduction

Face recognition might be one of the most well-known and attracting applications in computer vision. Because of the outstanding face recognition capability of human being and the numerous potential applications of face recognition technology, face recognition technology has been receiving more and more attentions. Over the past decades, researchers have focused on developing fully automatic facial recognition systems, which means, not only the recognition process but also face detection and feature extraction are all supposed to be handled by the computer automatically. As a result, a practical facial recognition system should at least include the three aforementioned parts.