Unifying Approach of Blind Image Restoration and Face Recognition in PCA Subspace

The performance of face recognition algorithms is reduced due to many challenging image acquisition conditions such as image blur and low resolution. This paper presents a unified approach for face image restoration and recognition. The idea for this unification is to solve two issues of face recognition and restoration using subspace method at the same time. The most popular subspace face recognition methods are based on PCA in which space face images can be efficiently represented as a lower-dimensional linear combination of basis images in terms of the variance maximization and independence respectively. In the context of subspace recognition scheme, the recent PCA-based image restoration approach is suitable to be unified with the restoration algorithm that can be used for face recognition method in PCA subspace. Some experiments are conducted to verify the effectiveness of the unified approach even in severe blurring image acquisition.

Keywords— Face recognition, Eigenface images, PCA subspace, Image restoration