Abstract:

An improved sparsity subspace learning method is proposed based on sparse representation. Firstly, the sparsity neighboring correlation reconstruction model is proposed to solve the insufficient data descriptions caused by global features in traditional sparsity subspace learning. Specifically, discriminative information is extracted from local structural information of all samples and label information of partial samples. Secondly, semisupervised technology is introduced to enhance the performance of subspace learning based on the sparsity neighboring correlation reconstruction method. Specifically, the regularization parameter is adopted to fuse label features of partial samples and sparsity neighboring reconstructive features of all samples. Experimental results demonstrate that the proposed method has good classification performance and recognition accuracy. Moreover, the experimental results also show that the sparsity neighboring reconstruction has good stability when extracting discriminative information.