Robust eye detection via sparse representation

As one of the distinct features in human faces, eyes play an important role in face alignment, face recognition and facial expression analysis. In this paper, we first study the sparse representation of an eye. Then a robust eye detection algorithm is proposed, which regards eye detection as a typical two-class classification problem and detects eyes based on the residual error of sparse representation. The performance of the proposed algorithm is subsequently validated using FRGC 1.0 database. The result shows that our eye detector has an overall 97% eye detection rate, which is very close to the manually provided eye positions.