Sparsity Preserving Canonical Correlation Analysis with Application in Feature Fusion
稀疏保持典型相关分析及在特征融合中的应用

Sparsity preserving projections (SPP) aim to preserve the sparse reconstructive relationship among the data and have been successfully applied in face recognition. The projections are invariant to rotations, rescalings, and translations of the data, and more importantly, they contain natural discriminating information even without class labels. Enlightened by this, we propose a sparsity preserving canonical correlation analysis (SPCCA). It can not only fuse the discriminative information of two feature sets efficiently, but also constrains the sparse reconstructive relationship among each feature set in order to increase the representational power and has good discrimination capability of the feature extracted by SPCCA. Experimental results on the multiple feature databases and face databases show that the proposed SPCCA is better than CCA.