On the Dimensionality Reduction for Sparse Representation based Face Recognition



In this project, we discussed the dimensionality reduction (DR) of face images when using sparse representation based classifier (SRC) for classification. Our experiments on Extended Yale B, AR and ORL face databases demonstrated that the proposed DR algorithm has better performance than Eigenfaces and Randomfaces. It can achieve higher recognition rate under the same dimensionality than Eigenfaces and Randomfaces. The proposed DR algorithm is an unsupervised learning method. In the future, we will investigate how to introduce the class label information into the DR learning so that a supervised DR algorithm can be developed to learn a set of discriminative projection matrix under the framework of SRC.

ref :

Zhang, L., Yang, M., Feng, Z., & Zhang, D. (2010, August). On the dimensionality reduction for sparse representation based face recognition. In Pattern Recognition (ICPR), 2010 20th International Conference on (pp. 1237-1240). IEEE.


There are no reviews yet.

Be the first to review “On the Dimensionality Reduction for Sparse Representation based Face Recognition”