Adjusting samples for obtaining better l2-norm minimization based sparse representation

• We evaluated imperfection of the state-of-the-art sparse representation based face recognition methods.
• We propose a novel l2-SR identifies the training sample that are important in correctly classifying samples.
• We provide in-depth analysis of l2-SR and provide new ideas to improve previous methods.

l2-norm sparse representation (l2-SR) based face recognition method has attracted increasing attention due to its excellent performance, simple algorithm and high computational efficiency. However, one of the drawbacks of l2-SR is that the test sample may be conspicuous difference from the training samples even from the same class and thus the method shows poor robustness. Another drawback is that l2-SR does not perform well in identifying the training samples that are trivial in correctly classifying the test sample. In this paper, to avoid the above imperfection, we proposed a novel l2-SR. We first identifies the training samples that are important in correctly classifying the test sample and then neglects components that cannot be represented by the training samples. The proposed method also involve in-depth analysis of l2-SR and provide novel ideas to improve previous methods. Experimental results on face datasets show that the proposed method can greatly improve l2-SR.